[House Hearing, 112 Congress]
[From the U.S. Government Publishing Office]
UTILIZING DNA TECHNOLOGY TO SOLVE
COLD CASES ACT OF 2011
=======================================================================
HEARING
BEFORE THE
SUBCOMMITTEE ON CRIME, TERRORISM,
AND HOMELAND SECURITY
OF THE
COMMITTEE ON THE JUDICIARY
HOUSE OF REPRESENTATIVES
ONE HUNDRED TWELFTH CONGRESS
SECOND SESSION
ON
H.R. 3361
__________
APRIL 25, 2012
__________
Serial No. 112-147
__________
Printed for the use of the Committee on the Judiciary
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Available via the World Wide Web: http://judiciary.house.gov
_____
U.S. GOVERNMENT PRINTING OFFICE
73-963 PDF WASHINGTON : 2012
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; DC
area (202) 512-1800 Fax: (202) 512-2104 Mail: Stop IDCC, Washington, DC
20402-0001
COMMITTEE ON THE JUDICIARY
LAMAR SMITH, Texas, Chairman
F. JAMES SENSENBRENNER, Jr., JOHN CONYERS, Jr., Michigan
Wisconsin HOWARD L. BERMAN, California
HOWARD COBLE, North Carolina JERROLD NADLER, New York
ELTON GALLEGLY, California ROBERT C. ``BOBBY'' SCOTT,
BOB GOODLATTE, Virginia Virginia
DANIEL E. LUNGREN, California MELVIN L. WATT, North Carolina
STEVE CHABOT, Ohio ZOE LOFGREN, California
DARRELL E. ISSA, California SHEILA JACKSON LEE, Texas
MIKE PENCE, Indiana MAXINE WATERS, California
J. RANDY FORBES, Virginia STEVE COHEN, Tennessee
STEVE KING, Iowa HENRY C. ``HANK'' JOHNSON, Jr.,
TRENT FRANKS, Arizona Georgia
LOUIE GOHMERT, Texas PEDRO R. PIERLUISI, Puerto Rico
JIM JORDAN, Ohio MIKE QUIGLEY, Illinois
TED POE, Texas JUDY CHU, California
JASON CHAFFETZ, Utah TED DEUTCH, Florida
TIM GRIFFIN, Arkansas LINDA T. SANCHEZ, California
TOM MARINO, Pennsylvania JARED POLIS, Colorado
TREY GOWDY, South Carolina
DENNIS ROSS, Florida
SANDY ADAMS, Florida
BEN QUAYLE, Arizona
MARK AMODEI, Nevada
Richard Hertling, Staff Director and Chief Counsel
Perry Apelbaum, Minority Staff Director and Chief Counsel
------
Subcommittee on Crime, Terrorism, and Homeland Security
F. JAMES SENSENBRENNER, Jr., Wisconsin, Chairman
LOUIE GOHMERT, Texas, Vice-Chairman
BOB GOODLATTE, Virginia ROBERT C. ``BOBBY'' SCOTT,
DANIEL E. LUNGREN, California Virginia
J. RANDY FORBES, Virginia STEVE COHEN, Tennessee
TED POE, Texas HENRY C. ``HANK'' JOHNSON, Jr.,
JASON CHAFFETZ, Utah Georgia
TIM GRIFFIN, Arkansas PEDRO R. PIERLUISI, Puerto Rico
TOM MARINO, Pennsylvania JUDY CHU, California
TREY GOWDY, South Carolina TED DEUTCH, Florida
SANDY ADAMS, Florida SHEILA JACKSON LEE, Texas
MARK AMODEI, Nevada MIKE QUIGLEY, Illinois
JARED POLIS, Colorado
Caroline Lynch, Chief Counsel
Bobby Vassar, Minority Counsel
C O N T E N T S
----------
APRIL 25, 2012
Page
THE BILL
H.R. 3361, the ``Utilizing DNA Technology to Solve Cold Cases Act
of 2011''...................................................... 4
OPENING STATEMENTS
The Honorable F. James Sensenbrenner, Jr., a Representative in
Congress from the State of Wisconsin, and Chairman,
Subcommittee on Crime, Terrorism, and Homeland Security........ 1
The Honorable Robert C. ``Bobby'' Scott, a Representative in
Congress from the State of Virginia, and Ranking Member,
Subcommittee on Crime, Terrorism, and Homeland Security........ 11
The Honorable John Conyers, Jr., a Representative in Congress
from the State of Michigan, and Ranking Member, Committee on
the Judiciary.................................................. 12
WITNESSES
Dennis P. Kilcoyne, Detective, Robbery and Homicide Division, Los
Angeles Police Department
Oral Testimony................................................. 15
Prepared Statement............................................. 17
Peter M. Marone, Director, Virginia Department of Forensic
Science
Oral Testimony................................................. 18
Prepared Statement............................................. 21
Henry T. Greely, Deane F. and Kate Edelman Johnson Professor of
Law, Stanford Law School
Oral Testimony................................................. 25
Prepared Statement............................................. 26
Michael T. Risher, Staff Attorney, American Civil Liberties Union
Foundation of Northern California
Oral Testimony................................................. 35
Prepared Statement............................................. 37
LETTERS, STATEMENTS, ETC., SUBMITTED FOR THE HEARING
Prepared Statement of the Honorable John Conyers, Jr., a
Representative in Congress from the State of Michigan, and
Ranking Member, Committee on the Judiciary..................... 13
UTILIZING DNA TECHNOLOGY TO SOLVE COLD CASES ACT OF 2011
----------
WEDNESDAY, APRIL 25, 2012
House of Representatives,
Subcommittee on Crime, Terrorism,
and Homeland Security,
Committee on the Judiciary,
Washington, DC.
The Subcommittee met, pursuant to call, at 10 a.m., in room
2141, Rayburn House Office Building, the Honorable F. James
Sensenbrenner, Jr. (Chairman of the Subcommittee) presiding.
Present: Representatives Sensenbrenner, Goodlatte, Lungren,
Scott, Johnson, Chu, Quigley and Conyers.
Also Present: Representative Schiff.
Staff Present: (Majority) Sam Ramer, Counsel; Arthur
Radford Baker, Counsel; Lindsay Hamilton, Clerk; (Minority)
Bobby Vassar, Subcommittee Chief Counsel; Ashley McDonald,
Counsel; and Veronica Eligan, Professional Staff Member.
Mr. Sensenbrenner. The Subcommittee will come to order.
Welcome everyone to today's hearing on H.R. 3361, the Utilizing
DNA Technology to Solve Cold Cases Act of 2011. It is hard to
remember a time in law enforcement before DNA searches became
commonplace. In 1953, when Watson and Crick published their
Seminal paper on the fundamental building blocks for an
individuals entire genetic makeup, few could have imagined the
investigative potential that would be unleashed. Law
enforcement officials have since used this evidence to capture
criminals and enhance public safety. It may have used it to
exonerate the innocent.
DNA is a powerful tool for law enforcement investigations
because each person's DNA is different from that of every other
individual except for identical twins. It cannot escape our
notice that in the past 20 years, there has been a dramatic
decrease in crime levels across the United States. According to
the FBI, violent crime in the U.S. have dropped by almost 50
percent in the last 20 years. DNA is a part of new arsenal of
tools that may have contributed to the capture and imprisonment
of repeat offenders, and may have enhanced the Nation's safety.
Federal law authorizes the FBI to operate and maintain a
National DNA database called CODIS where our DNA profiles
generated from samples collected from people under applicable
legal authority and samples collected at crime scenes can be
compared to generate leads in criminal investigations.
Statutory provisions also authorize the collection of DNA
samples from Federal offenders and arrestees, D.C. offenders
and military offenders. State laws dictate which convicted
offenders. And sometimes people arrested for crimes have will
have profiles entered into state DNA databases while Federal
law dictates the scope of the national database.
Some jurisdictions have started to use their DNA databases
for familial searching. Familial DNA searches scan the database
for individuals related by some degree to the target suspect.
By broadening the search parameters of the DNA code,
investigators can find the siblings, children or more distant
relatives of an individual. Some jurisdictions have started to
use their individual State DNA databases for familial
searching. So far two States, Colorado and California make the
most use of familial DNA.
In March 2011 Governor McDonnell announced that the
Virginia Department of Forensic Science would begin using
familial DNA searches in Virginia. Several States, however,
including Alaska, Maine, Michigan and Vermont include
prohibitions on either partial, match or familial match
searching in lab manuals. The FBI currently does not permit
familial DNA searches of the CODIS databank.
In a recent publicized case in California, a serial killer
suspect in Los Angeles was identified using this method. The
police called him the Grim Sleeper because he seemed to go
dormant in between murdering at least 10 women over more than
20 years. Saddled with an ice cold case, California authorities
decided to run DNA samples from saliva left on the bodies
through a familial DNA search. They identified a young man
named Christopher who was in the system serving a prison
sentence on a felony weapons charge.
Further investigation led police to 57-year-old Lonnie
David Franklin, Christopher's father. Once police had him as a
suspect, an undercover detective posing as a waiter collected
the older Franklin's plate, utensils and leftover food from the
restaurant, and the suspect's DNA was found on a discarded
pizza crust matched the DNA left long ago on the bodies of the
dead women. The case suddenly came back to life. Franklin is
now awaiting trial on multiple charges of murder. This case has
brought significant attention to the technique of familial DNA
searches.
H.R. 3361 utilizing DNA Technology to Solve Cold Cases Act
requires the Attorney General to adopt policies and procedures
to permit the FBI to conduct familial searches for DNA samples
collected from crime scenes and Federal investigations, and
that a State administrator or State Attorney General may
request that the FBI conduct familial searches for DNA samples
collected from crime scenes and State investigations, and that
the privacy interest of persons identified in familial searches
are carefully protected.
The Act imposes restrictions on such familial DNA searches.
For example, it limits the familial search to cases where there
is no identical match from the crime scene with someone in the
offender database. In addition, the bill limits a familial DNA
search to the investigation of the following crimes: First, an
offense of murder, voluntary manslaughter, kidnapping or any
attempt to commit murder, voluntary manslaughter, or
kidnapping. Second, a specified offense against a minor or an
attempt to commit such a specified offense, or third, an
offense for which an offender would be required under the Sex
Offender Registration and Notification Act, 42 U.S.C. 16-901,
and seek to register as a tier 3 sex offender, or an attempt to
commit such an offense.
In addition, the States are required to have written
policies in place that are consistent with those of the
Attorney General. The bill also has reporting requirements to
the House and Senate Judiciary Committee so that the use of
familial searches would have significant oversight.
Today we will look at familial DNA searches looking at the
efficacy of the technique, the ethical and privacy issues that
it may entail. Modern 21st century law enforcement has improved
and reduced many of the problems we were concerned about back
in the 20th century. As the crime rate continues to fall we
should make sure that our attack has continued to evolve and
target criminals with continued respect for our citizens
privacy, interest and civil liberties so that we may protect
Americans without sacrificing the values that we all hold
sacred.
I looked forward to hearing more about this issue and thank
all of our witnesses for participating in today's hearing. And
now I recognize the gentleman from Virginia, Mr. Scott for an
opening statement.
[The bill, H.R. 3361, follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
__________
Mr. Scott. Thank you, Mr. Chairman. I am pleased to join
you in this hearing on the familial DNA testing. I think it is
wise that we have a hearing about this bill and the issue
before moving legislation on it. DNA is a powerful enforcement
tool, but when not used carefully and with proper procedures in
place, it can do great harm to privacy and other
constitutionally-protected rights. As I expect to hear from our
witnesses, familial DNA searching differs from traditional DNA
searching in that law enforcement is not seeking from the
database one direct match, but rather, a match from a familial
search is actually only an investigatory tool because it is a
close match, not an exact match. It directs law enforcement to
a person who is not a perpetrator, but who is said likely to be
a relative.
For example, for a given crime the crime scene sample
matches no one in the database, partially matches person X who
is in the database because he has been arrested for--was once
arrested for a felony.
If person X has three brothers, a father and two sons, all
six of those relatives, or possibly, five out of six of those
relatives are actually innocent of the crime, but those five
relatives will now be--all six relatives now subject to police
investigation and including questioning, request for DNA
samples, and/or surreptitious collection of DNA samples merely
because they are related to someone whose DNA looks a lot like
but is not identical to the DNA found at a crime scene.
This way, the search leads investigators to the doors of
many people who are, in fact, innocent of the crime, but only
come under suspicion by happenstance of being related to
someone who is in the DNA database. This investigative method,
if used, will undoubtedly resolve in apprehending some
additional perpetrators who might not have otherwise been
caught, just as it happened in the Grim Sleeper case. But
before we rush to authorize this kind of testing, we need to
closely examine the societal costs and the societal harms.
One of the costs is invasion of privacy I alluded to
earlier, a familial DNA searching involves identifying and
investigating a group of people, all of whom or all but one of
whom are innocent of unconnected to the crime being
investigated. One of the primary privacy concerns regarding the
use of this kind of testing is that it will put innocent people
under genetic surveillance because they are related to someone
whose profile is in the DNA database.
Another societal cost is the impact it will have on
minority communities, African-American and Hispanics are
disproportionately represented in the DNA index system because
they are disproportionately arrested and convicted. Now, that
is not always related to the incidents of criminal activity,
because we have found that African-American represent 40
percent of the drug arrests compared to 13 percent of the
general population where there is no reason to believe that
Blacks are actually committing more drug offenses than Whites.
Third, the societal cost is the number of false positives
that will result. False positives include both long lists of
partial matches that could be investigated, again, any number
of relatives for each partial match who will be investigated.
These false positives will be financially costly for law
enforcement and will further increase privacy concerns and will
impact minority communities disproportionately.
I, therefore, have grave concerns about the societal costs
of this kind of testing. The only way to completely eliminate
the problems associated with that would be to prohibit it
altogether, as Maryland and District of Columbia have done. And
if a familial DNA testing is to be conducted, which should be
done on a State-by-State basis among the States and not as a
national or Federal program. DNA continues to evolve as a
technology, and may some day evolve to the point of great
enough certainty and sufficient safeguards to justify its use.
I am concerned we are not quite at that point yet. I look
forward to the testimony of our witnesses on whether technology
is and whether it is sufficient enough to be developed to go
national or Federal at this time.
Mr. Chairman, I would like to welcome one of the witnesses
Pete Marone from Virginia. Virginia has a long history of being
in the forefront of DNA technology and the Virginia forensic
lab is one of the best and it is great to have him here. I also
want to recognize the chief sponsor of the legislation we are
considering, a gentleman from California, Mr. Schiff, a former
Member of this Committee and a former prosecutor. I yield back.
Mr. Sensenbrenner. The gentleman from Michigan, Mr.
Conyers, the Ranking Member of the full Committee is recognized
for an opening statement.
Mr. Conyers. Thank you, Chairman Sensenbrenner. I join in
welcoming Adam Schiff for the discussion today. I wish he was
back on the Committee.
I agree with all the comments that the Chairman and the
Ranking Member have made, so I will put my statement in the
record and just add these couple comments. It seems to me, at
first blush for this discussion, that the cost outweigh the
benefits, and that we need more safeguards. Attorney Michael
Risher will probably expand on that, representing the American
Civil Liberties Union. And I had an opportunity to talk with
the Wayne County prosecutor of Michigan, Ms. Kym Worthy, about
this hearing this morning. And she was telling me about the
backlog of 11,000 untested rape kits that are on their shelves
right now. Some of them clearly unusable because of the
expiration of the statute of limitations. These kits, to my
surprise, cost $1,500 each. And they have only funding for a
very small number of them.
So we have a problem here, as Bobby Scott has already
indicated, that African-American and Hispanics are already
overrepresented in the national DNA index system, because they
are disproportionately arrested and convicted. And it seems
inescapable to me, and our distinguished witnesses can give me
their view on this, that this disparity will probably go up the
more we use this system.
So I merely wanted us all to appreciate that this may be an
idea that we will have to examine far more carefully than
merely one case. It involves a serious problem of expanding the
invasiveness of the criminal justice system. And by the way,
this hearing comes along at a very appropriate time for me
because I was just reading recently of a former United States
Supreme Court Justice that was talking about the breakdown of
the criminal justice system in our country, and the problems
that are attendant with that, and this could be moving us in
the wrong direction.
I want to commend the Subcommittee for taking this up, and
I hope that the two leaders of the Subcommittee will allow us
to begin to make some other inquiries about the criminal
justice system in general and in particular, as the session
moves on.
Thank you very much, Mr. Sensenbrenner.
[The prepared statement of Mr. Conyers follows:]
Prepared Statement of the Honorable John Conyers, Jr., a Representative
in Congress from the State of Michigan, and Ranking Member, Committee
on the Judiciary
In the past decade, DNA technology has become increasingly vital to
ensuring accuracy and fairness in the criminal justice system. Where
biological evidence exists in an unsolved ``cold case,'' such evidence
can be a powerful tool to help investigators solve the crime. DNA can
also be used to clear suspects and exonerate people wrongly accused or
convicted of crimes.
Because DNA is such a powerful tool, however, we must be careful
how we use it. H.R. 3361 deals with familial DNA searching. A familial
search refers to searching in a DNA database not for the person who
left the DNA sample at the crime scene but for a relative of that
person.
In familial DNA searching, law enforcement is no longer seeking
from the database one direct match between a crime scene sample and a
perpetrator's sample. The ``match'' that results from a familial search
is merely an investigative lead that directs law enforcement to a
person who is not the perpetrator but who may be a relative of the
perpetrator.
This is a dramatic expansion of a traditional DNA search. Before we
rush to authorize familial DNA testing, we need to closely examine the
individual, familial, and societal costs and harms.
FIRST, minority communities are disproportionately impacted by
familial searching. African Americans and Hispanics are over-
represented in the National DNA Index System (or ``NDIS'') because they
are disproportionately arrested and convicted. Professor Henry Greeley
has estimated that using the NDIS for familial searching could mean
that approximately 17% of the African American population in the United
States would be ``findable'' through the database, compared to
approximately 4% of the white population. This means that, by far, the
majority of the innocent people who will be affected by familial
searching will be African American and Hispanic.
SECOND, a high number of false positives will result from familial
testing. False positives include both long lists of partial matches
that could be investigated by local law enforcement and also many
relatives for each partial match who could be investigated. These false
positives will be financially costly for local law enforcement to
investigate. This takes resources away from other important law
enforcement activities. Kym Worthy, the prosecutor of Wayne County,
Michigan, informs me that there is a backlog of 11,000 untested DNA
rape kits in her county. She tells me it costs approximately $1500 per
rape kit to test. If she had the funding, she would test all of those
rape kits. Diverting scarce resources to familial DNA searches means
fewer resources for other important law enforcement activities.
FINALLY, familial searching invades the privacy of innocent people.
Familial searching subjects relatives of convicted offenders to
potential law enforcement scrutiny, without probable cause, and puts
innocent people under ``genetic surveillance'' merely because they are
related to someone whose profile is in a DNA database.
I therefore have tremendous concerns about the societal costs of
familial DNA searching. These individual, familial, and societal
concerns must be balanced against the benefits when we decide whether
to enact legislation that will permit familial searching on a federal
level. I believe the costs of familial DNA searching might outweigh the
benefits, and we need more safeguards if we are going to enact federal
legislation.
If we decide that, on balance, we want a federal law permitting
familial searching, we still must ensure that there are adequate
safeguards in place, such as requirements that all other investigative
leads have been exhausted and that any familial searching only be
permitted to help solve major violent crimes where there is a
continuing and serious risk to public safety. The witnesses testifying
today will address some of the other safeguards that need to be in
place and that H.R. 3361 lacks.
I applaud Chairman Sensenbrenner for calling a hearing today about
this bill and this is issue that merits further, careful study before
we act. I yield back.
__________
Mr. Sensenbrenner. I thank the distinguished Ranking Member
for his statement. Without objection, all Members opening
statements will appear in the record. It is now my pleasure to
introduce today's witnesses: Detective Dennis Kilcoyne is a
supervising detective for the Los Angeles Police Department. He
has been a member of the LAPD for over 35 years. Twenty-seven
of those years have been dedicated to investigating homicides
and major crimes. Since 1994, Detective Kilcoyne has worked for
the LAPD's elite robbery homicide division. He currently serves
as the president of the California Homicide Investigators
Association, a position he has held for the past 8 years. This
association is made up of local law enforcement investigators,
prosecutors and death investigation professionals from
throughout California.
Peter Marone has been the director of the Virginia
Department of Forensic Science since 2007. He has been with the
Department since 2005, previously serving as the director of
technical services. From 1998 to 2005, Mr. Marone was the
central laboratory director at the Division of Forensic
Science. And for 1986 to 1998 he previously held the position
of assistant division director and program manager.
Previously, he had been a forensic scientist at the
Virginia Bureau of Forensic Science and a criminologist at the
Allegheny County Crime Laboratory in Pittsburgh. He received
his bachelor of science in chemistry in 1970, and his Masters
in for forensic chemistry in 1971 from the University of
Pittsburgh.
Henry Greely is the Deane F. and Kate Edelman Johnson
Professor of law and professor by courtesy of genetics at
Stanford. He chairs the California Advisory Committee on Human
Stem Cell Research and steering committee of the Stanford
University Center for Biomedical Ethics, and directs the
Stanford law or Stanford Center for Law and biosciences. Before
starting his career at Stanford in 1985, Professor Greely
served as a law clerk for Judge John Minor Wisdom of the U.S.
Court of Appeals and for Justice Potter Stewart of the U.S.
Supreme Court.
After working during the Carter administration in the
Departments of Defense and Energy, he entered private practice
in Los Angeles in 1981 as a litigator with the law firm of
Tuttle & Taylor, Inc. He graduated from Stanford in 1974 and
from Yale Law School in 1977.
Michael T. Risher is a staff attorney for the ACLU in
northern California, the Nation's largest ACLU affiliate.
Before joining the ACLU NC, Mr. Risher was a deputy public
defender in Alameda County from 1998 to 2005. He also served as
the legal affairs advisor for the Linda Smith Center and a
clerk to Judge Karen Nelson Moore of the U.S. Court of Appeals.
He is a graduate of Harvard and Stanford Law School.
The witnesses written statements will be entered into the
record in their entirety. I ask that you summarize your
statements in 5 minutes or less. You all know about the green,
yellow and red, and the Chair has the big gavel. Detective
Kilcoyne, you are first.
TESTIMONY OF DENNIS P. KILCOYNE, DETECTIVE, ROBBERY AND
HOMICIDE DIVISION, LOS ANGELES POLICE DEPARTMENT
Mr. Kilcoyne. Thank you, Mr. Sensenbrenner and Committee
Members for allowing Los Angeles Police Department to comment
on this bill. In May of 2007, detectives from Los Angeles
Police Department's robbery homicide division received
information from the forensic lab regarding case-to-case hits,
which are linked by DNA matches to two LAPD murders as well as
2002 Inglewood murder of a 14-year-old girl. All three cases
involved young women and were unsolved. Biological evidence
returned to one individual, however, his identity was absent
from any databank.
The Department established a task force to investigate the
series of crimes, and within the first months of research into
years of cold cases, a similar series involving nine cases
between 1985 and 1988 were connected to the current series. One
of the nine cases included a surviving victim, who 25 years
earlier, had been sexually assaulted, photographed and left for
dead by a lone male gunman. During the 1980's, a 200-member
task force had investigated these heinous crimes and not been
able to identify the suspect. DNA, as an investigative tool,
had yet to be developed for law enforcement at that time.
The task force renewed the effort to identify and apprehend
the suspect, and there was widespread media attention and
public outreach campaign for information that led to over a
1,000 tips provided by the community. For the next 2 years,
detectives pursued leads all over the Nation. Sadly, detectives
were no closer to identifying the suspect than the original
detectives were decades earlier. During the summer of 2008,
detectives developed a partnership with the California
Department of Justice Bureau of Forensic Services, regarding
the Bureau's development of new software to search California's
convicted felon databank for matches that have a familial
genetic connection to the crime scene evidence.
The creation of this program was based on a series of
crimes that LAPD was investigating and was the model for its
uses. A strict protocol was established by the Department of
Justice to set guidelines for the usage of a familial search.
Case consideration must meet the following: Number 1, there
must be a crime of violence that includes critical public
safety implications; number 2, all reasonable and viable
investigative leads have been exhausted, and the biological
evidence is from a single source profile exhibiting a minimum
of 15 genetic markers.
The requesting investigative agency, prosecutor and DOJ,
then enter into a signed memorandum of understanding. All
requests, analysis results and disclosure of findings are
handled by DOJ familial search committee. If a familial match
is found, the committee determines if the information warrants
further inquiry. The information is then investigated by DOJ's
Bureau of Investigation using public databanks to verify the
findings through State identification, birth records, property
records, et cetera.
This information is then presented to the committee for
additional review, all of this review takes place without the
knowledge or communication with the requesting police agency or
prosecutor. When the familial connection is verified and
approved, a formal meeting is called with the lead law
enforcement agency investigators and prosecutors. In the case
of the series I have described, the information was only shared
with me and the Chief of Police, Charlie Beck.
The next step is to conduct surveillance on the suspect and
obtain publicly discarded items containing DNA. Such items are
submitted to the forensic lab for analysis. When a match
between the DNA sample and the individual is made, probable
cause has been established for detention. After the suspect is
detained, a court order confirmation DNA swab is obtained
directly from the suspect and is confirmed as a direct match to
the crime scene evidence prior to formal charges being filed.
In November of 2008, the first familial search run was done
with the eyes of the forensic world watching. Unfortunately, no
match was made at this time. The detective work continued for
another year and a half including renewal of reward officers,
billboard campaigns, and continued investigation of tips that
were pursued all over the country. With the passage of so much
time, investigators wondered if the perpetrator was still in
the country, or if he was even still alive.
A second formal request was made with the Cal DOJ in the
spring of 2010. Detectives with the DOJ forensic chief opined
the databank pool had grown over time and offered more
opportunity for a match. The tide turned in June of 2010, the
second search of a convicted felon databank produced a match to
the son of Lonnie David Franklin. The son had been recently
been convicted for a felony crime and his DNA sample had been
obtained in accordance with the DNA collection log. Franklin,
the father, was a former city employee who resided in the heart
of Los Angeles. Franklin was immediately put under
surveillance.
Mr. Sensenbrenner. Detective, your time has expired, could
you wrap it up, please.
Mr. Kilcoyne. Yes, sir. Franklin--DNA was collected, he was
a match to the case to the direct evidence collected at
multiple crime scenes. Familial DNA is certainly worthy of
discussion and uniform control, strict guidelines such as those
in place in California must be followed to ensure careful
review of evidence, adherence to scientific protocol,
consideration of collection sample regulations, privacy issue
and protection of the innocent and the apprehension of the
guilty.
Mr. Chairman and Members of the Subcommittee, thank you for
inviting me to speak today. I am now ready to take any
questions you may have.
[The prepared statement of Mr. Kilcoyne follows:]
Prepared Statement of Dennis P. Kilcoyne, Detective,
Los Angeles Police Department
Chairman Smith, Ranking Member Conyers, and distinguished members
of the Committee, thank you for the opportunity to discuss the Los
Angeles Police Department's (LAPD) view and insight as to the value of
utilizing DNA technology to solve cold cases.
In May 2007, detectives at the Los Angeles Police Department
Robbery-Homicide Division received information from the Forensic lab
regarding ``case to case hits,'' by DNA matches, to two LAPD murders in
2007 and 2003 as well as a 2002 Inglewood murder of a 14-year-old girl.
All three cases involved young women and were unsolved. Biological
evidence returned to one individual, however his identity was absent
from any databank.
The Department established a task force to investigate this series
of crimes and within the first months of research into years of cold
cases, a similar series involving nine cases between 1985 and 1988 were
connected to the current series. One of the nine cases included a
surviving victim who 25 years earlier had been sexually assaulted,
photographed, shot and left for dead by a lone male gunman. During the
1980's, a 200 member task force had investigated these heinous crimes
and had not been able to identify a suspect. DNA as an investigative
tool had yet to be developed for law enforcement at that time.
The Task Force renewed the effort to identify and apprehend the
suspect. There was widespread media attention and a public outreach
campaign for information that led to over 1,000 tips provided by the
community. For the next 2 years, detectives pursued leads all over the
nation. Sadly, the detectives were no closer to identifying the suspect
than the original detectives were decades earlier.
During the summer of 2008, detectives developed a partnership with
the California Department of Justice, Bureau of Forensic Services
regarding the bureau's development of new software to search
California's ``Convicted Felon databank'' for matches that have a
``familial genetic connection'' to the crime scene DNA evidence. The
creation of this program was based on the series of crimes that LAPD
was investigating and was the model for its usage.
A strict protocol was established by the Department of Justice
setting guidelines for the usage of a ``Familial Search''. Case
consideration must meet the following:
1. Must be a crime of violence and include critical public
safety implications.
2. All reasonable and viable investigative leads have been
exhausted.
3. The biological evidence is from a single source profile
exhibiting a minimum of 15 genetic markers (15 Short Tandem
Repeats (STR) loci (location on the genetic marker).
The requesting investigative agency, prosecutor and the DOJ then
enter into a signed Memorandum of Understanding (MOU). All requests,
analysis results and disclosure of findings are handled by a DOJ
Familial Search Committee. If a familial match is found, the committee
determines if the information warrants further inquiry. The information
is then investigated by the DOJ Bureau of Investigations using public
databanks to verify the findings through state identifications, birth
records, property records etc. This information is then presented to
the committee for additional revue. All of the review takes place
without the knowledge or communication with the requesting agency or
prosecutor. When the Familial connection is verified and approved, a
formal meeting is called with the lead law enforcement agency
investigators and prosecutors. In the case of the series I have
described, the information was only shared with me and the Chief of
Police, Charlie Beck.
The next step is to conduct surveillance on the suspect and obtain
a publicly discarded item containing DNA. Such items are submitted to
the forensic lab for analysis. When a match between the DNA sample and
an individual is made, Probable Cause has been established for a
detention. After the suspect is detained a Court ordered confirmation
DNA swab is obtained directly from the suspect and confirmed as a
direct match to the crime scene evidence prior to formal charges being
filed.
In November of 2008 the first familial search run was done with the
eyes of the forensic world watching. Unfortunately, no match was made
at that time. The detective work continued for another year and a half
and included renewal of reward offers, billboard campaigns, and
continued investigation of tips that again were pursued all over the
country. With the passage of so much time, investigators wondered if
the perpetrator was still in the country, or if he was even still
alive.
A second formal request was made with the California Department of
Justice in the spring of 2010. Detectives and the DOJ Forensic chief
opined the data bank pool had grown over time and offered more
opportunity for a match. The tide turned on June 30, 2010. The second
search of the convicted felon databank produced a match to the son of
Lonnie David Franklin. The son had recently been convicted for a felony
crime and his DNA sample had been obtained in accordance with a DNA
collection law. Franklin, the father, was a former city employee who
had resided in the heart of South Los Angeles during this most prolific
series of violent crime in Los Angeles history.
Franklin was immediately put under surveillance as a sample of his
DNA was needed to confirm a match. At a local restaurant, a discarded
pizza crust, collected by a detective posing as a waiter yielded a DNA
match to the DNA left by the suspect in the multiple murders. Franklin
remains in custody and is awaiting trial in Los Angeles, charged with
10 murders and one attempted murder.
Since his arrest detectives have linked seven additional cases to
Franklin. The violence that went on for so long is the best argument I
can think of that modern law enforcement must have forensic advances as
tools to prevent and stop this type of terror in our communities.
The Familial DNA arena is certainly worthy of discussion and
uniform control. Strict guidelines, such as those in place in
California must be followed to ensure careful review of the evidence,
adherence to scientific protocol, consideration of collection sample
regulations, privacy issues, protection of the innocent and
apprehension of the guilty. The advancement of science utilized to
protect the public should be viewed as a tool that makes us all safer.
Mr. Chairman and members of the subcommittee, thank you for
inviting me to speak today. I am now ready to answer any questions you
may have.
__________
Mr. Sensenbrenner. Mr. Marone.
Mr. Marone. Mr. Chairman.
Mr. Sensenbrenner. Please turn your mic on. I don't think
it is on.
TESTIMONY OF PETER M. MARONE, DIRECTOR,
VIRGINIA DEPARTMENT OF FORENSIC SCIENCE
Mr. Marone. Thank you for inviting me to speak. I am the
director of the Department of Forensic Science. The issue I
have been specifically requested to speak on is familial
searching. Although Virginia official began familial searching
in April of 2011, that date was preceded by substantial,
technical, and more importantly, policy and logistical
discussions. While I strongly support the use of familial
searching as a means of identifying perpetrators of serious
crimes against a person after all investigative leads have been
exhausted. Having said that, I feel it is important it for all
to understand that many of these logistical as well as
technical issues are involved to be able to implement the
program.
Familial searching is an intentional or deliberate search
of a DNA database designed to identify relatives of offenders
as possible perpetrators. It is inherited by members of a
family, children will inherit half from mom, half from dad. And
siblings, consequently, will tend to share a larger portion of
the DNA types than unrelated people.
Under the current procedures, a typical search of a
database that results in a databank hit means that there is a
match between the crime scene evidence and an offender's
sample. The match is defined in this example is the DNA profile
from evidence or the types are identical to the DNA types from
the individual, the offender or arrestee.
When there is not a match between the evidence and
offender, a search of the database can be conducted to
determine if an individual has a similar, but not matching
profile. I will add at this point is you are not using the same
software as the CODIS software. It is an entirely different
module that would have to be utilized to do that.
A search such as this would yield a multiple candidate
because the search requirements are less stringent. When you
are doing a familial search, you are looking for close matches
or partial matches if you will. In order to get that number, to
get a reasonable number, we have to look at a window of
inclusion. You need those numbers to be small enough, but not
too small that would exclude possible relatives. On the other
hand, it can't be too big that it pulls in too many people who
aren't related at all.
It is important to understand that in a larger database,
the larger the database is, the greater the number of potential
relatives are generated. These candidates who have been
identified by familial search may have a biological
relationship to the evidence, a sibling, a parent. What we are
talking about is direct relationship, father to son, son to
father or brothers. But it is more likely that none of the
candidates identified in a familial search will be relatives,
all but one, will be relatives of the individual who deposited
the crime scene. Just because you are in the number of
candidates produced doesn't mean that the individual will be
identified as a possible relative.
The national recommendations that are put forward by SWGDAM
States that if a laboratory decides to perform familial
searches--and by the way, there is no forbidding in Federal law
to perform DNA searches, or familial searches. They should
generally be conducted on DNA profiles of a single source and
not a mixture. If you throw in mixtures, you have multiple
people and it is impossible to come up with a meaningful result
from there.
Since the purpose of the current databank search software
is to identify only those individuals whose profile exactly
matches, as I said before, the alternative software has to be
developed or purchased and there are software packages
available for this purpose.
A familial search profile is conducted in a DNA databank
looking for similar profiles. One approach is to then rank the
return candidates statistically to determine how likely they
are to be related to the person who deposited the biological
evidence. A ranking of these individuals conducted by computer
software and then top candidates are selected to give
additional DNA testing, the additional DNA testing is conducted
on the evidence and on the ranked candidates.
Lineage markers are used for the additional DNA testing,
these are DNA types passed on from father to son, within the
family. And one of the things I want to add is you are not
looking for Uncle Charlie or a few cousins, it is a direct
relationship. In California, they have a databank the size of
approximately 1.2 million, by now it might be 1.3 offenders. A
familial search of an evidence profile against the database
will generate many potential relatives, potential relatives;
the top candidates may be approximately 200 of them are
subjected to additional lineage testing as well as the evidence
to get that sample. That is why STRs, you are looking at the
male chromosome.
Since familial searching involves identifying and
investigating persons who are unconnected to the crime being
investigated, criteria must be established implementing the
procedure that balances the need against the use of resources
infringement on personal property.
I will cut to the chase by saying we utilize all the same
safeguards that California uses, and on a national basis, there
is one caveat to remember: The larger the database, the larger
the number of people involved.
A big problem with having a national search is that each
one of the samples of the State databases resides at the State
level. So currently, it would be very, very logistically
problematic for a Federal agency, the FBI to perform the Y STRs
on samples that are located in the States, that is one of the
things.
So what I propose is, just as the U.K. Does geographical
filtering, that the filtering be done at the State level, 85 to
90 percent of the hits that are made are at the State level,
they are not State to State. Secondarily to that would be maybe
contiguous States, but again you have to be careful that those
States don't have laws precluding that. In our case, it would
be Maryland, Maryland doesn't do familial testing.
Mr. Sensenbrenner. Thank you.
[The prepared statement of Mr. Marone follows:]
Prepared Statement of Peter M. Marone, Director,
Virginia Department of Forensic Science
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
__________
Mr. Sensenbrenner. Professor Greely.
TESTIMONY OF HENRY T. GREELY, DEANE F. AND KATE EDELMAN JOHNSON
PROFESSOR OF LAW, STANFORD LAW SCHOOL
Mr. Greely. Thank you, Mr. Chairman and Members of the
Committee. I greatly appreciate the honor of being here,
particularly as a law professor, and I can tell you the
students in my law and genetic seminar are very excited about
this hearing. I had the good luck in 2006 to be the lead author
on what I think is the first article to examine these issues
closely along, with three geneticists, Joanna Mountain, Daniel
Riordan, and Nanibaa Garrison. We viewed this as a mixed issue,
that were pluses and there were minuses. Six years later, I
think the issue remains the same. There are some real, but
limited benefits from this procedure. There are some real but
limited costs to this procedure. On balance, I support this
bill, but I do think it is not a panacea, neither is it a
monster.
In my brief time, I want to say three things: I want to
tell you a little bit about the scientific benefits and limits
of the procedure; a little bit about the policy and ethical
issues; and then about possible extensions and why having an
Attorney General a requirement that the Attorney General have
regulations here is, I think, a particularly good thing.
First, the greatest benefit of this procedure is it can
lead to clues that can solve cases that otherwise cannot be
solved, or have not been solved. There is now 10 years of
experience with that in the United Kingdom, there is limited
experience with it in the United States. The Grim Sleeper is, I
think, the most dramatic success but it is not the only
success. The biggest problem with this technique, at least
operating on the current CODIS markers is that it is not very
efficient.
We calculated in 2006 that a person with an averagely rare
or an averagely common set of CODIS markers would have
somewhere around 2,000 to 3,000 potential family member matches
in the CODIS database. And that is when the CODIS database was
one quarter of its current size. So someone with an average
genotype will throw up 8,000, 9,000, 10,000 hits. Now on one
hand, you can view that as a civil liberties disaster if all
10,000 of those people get interviewed and have DNA taken.
On the other hand, the police are not going to interview
10,000 people. It going to be a limited procedure because the
cost, most of the time it will throw up too many potential
leads, the leads will need to be winnowed down geographically
or through other DNA markers, or it will have to be in a case
that is of such importance that the investigators are willing
to put a great deal of time and effort into it. It will be a
contributor, it will not be a huge contributor.
The costs are also real but limited. There are some privacy
costs, there are some costs in terms of the inconvenience of
being a suspect. I haven't been a suspect since I was a
teenager and was pulled over, I didn't like it. I wouldn't like
it now, I think. One of the advantages of a DNA-based suspect
status, though, is that it is almost impossible for you to be
falsely convicted, or even falsely tried. If there is crime
scene DNA, and you are not a match for that crime scene DNA,
you will not be prosecuted and you will not be convicted.
However, the mere fact of being interviewed by the police is
perhaps not the most enjoyable of circumstances for anybody
involved.
The other cost, and to me back in 2006, and again today,
this is my biggest concern about it because of the ethnic
makeup of the CODIS database, this cost will disproportionately
fall on the African American community. That is not fair. On
the other hand, right now, the cost of investigations and the
cost of crimes fall disproportionately on the African American
community. You may be investigated for reasons other than DNA,
but the investigations will still focus the suspects--because
of the crime conviction rates, the suspects are likely to be
disproportionately African American today. At least this is a
technique that cannot lead to a false conviction. It is very,
very difficult for it to lead to a false conviction.
The last point I would make in the long run, this technique
can be made better by adding different genetic markers,
additional genetic markers to the 13 CODIS markers. But as my
written testimony points out, those raise some really
complicated issues of their own, that is part of what
California's doing with the Y chromosome is adding a different
set of markers. That leads to a better result, both for law
enforcement and for the public because there will be fewer
false positives, the results will be more useful and fewer
innocent people will be interrogated, but there are things that
need to be considered, so I think the idea of having the
Attorney General look at and make regulations here is quite
important.
On balance, I think this is a useful procedure, it is not a
panacea, it has costs, it is not a monster. It needs careful
scrutiny and regulation. This bill's provision for the Attorney
General to make regulations provides that. I hope you will
support the bill.
Mr. Sensenbrenner. Thank you, Professor Greely.
[The prepared statement of Mr. Greely follows:]
Prepared Statement of Henry T. Greely,*
Deane F. and Kate Edelman Johnson Professor Law, Stanford Law School
---------------------------------------------------------------------------
* Deane F. and Kate Edelman Johnson Professor of Law; Professor, by
courtesy, of Genetics; Director, Center for Law and the Biosciences;
Stanford University. The views expressed in this testimony are not
necessarily those of Stanford University--or anyone else. I would like
to thank my co-authors from our 2006 paper, Drs. Joanna Mountain,
Daniel Riordan, and Nanibaa Garrison.
---------------------------------------------------------------------------
``The sins of the fathers are to be laid upon the children.'' \1\
---------------------------------------------------------------------------
\1\ The Merchant of Venice, Act III, Scene 5, line 1.
This biblical-sounding quotation is actually from The Merchant of
Venice but what Shakespeare meant by it unclear, as he gives the line
to the play's fool. The Bible itself, at least in the King James
version, does not use exactly this language, but in at least five
places expresses similar sentiments about the Lord visiting the
``iniquity'' of the fathers on several generations of children.\2\ On
the other hand, Ezekiel states ``The son shall not bear the iniquity of
the father, neither shall the father bear the iniquity of the son: the
righteousness of the righteous shall be upon him, and the wickedness of
the wicked shall be upon him.'' \3\
---------------------------------------------------------------------------
\2\ See Exodus, 20:5, Exodus 32:7, Numbers 14:18, Deuteronomy 5:9,
and Jeremiah 32:18.
\3\ Ezekiel 18:20.
---------------------------------------------------------------------------
In recent years new uses for forensic DNA matching have provoked
similarly mixed reactions about the family connections and, perhaps not
sin or iniquity, but crime. Our now ``traditional,'' but, in fact, less
than 20 year old, forensic use of deoxyribonucleic acid (``DNA'')
compares DNA profiles from crime scene DNA to either the profiles of
particular suspects, or, through DNA databases, the profiles of people
convicted of crimes--and, increasing, of people arrested for felonies
or of non-U.S. nationals ``detained'' by the federal government. This
method looks for a perfect or near-perfect match, indicating that the
crime scene DNA almost certainly came from the suspect or from a person
in the database (or from his identical twin). Family forensic DNA is a
technique used when there is no perfect match, in the hope of
generating investigative leads by seeing whether the crime scene DNA is
likely to have come from a close genetic relative of a person in the
database.
I was part of a group that published one of the first close
analyses of family forensic DNA \4\, in 2006, and have continued to
follow the issue. I believed then, and continue to believe now, that
family forensic DNA, using our current technology, is a weak,
inefficient, but occasionally useful method for generating
investigative leads. I also believed, and continue to believe, that,
although its use is disquieting, it raises no strong constitutional or
other legal questions. It does raise a few policy problems, some, but
not all, of which can be mitigated by regulating its use. Although it
is not a panacea, the federal government should allow its careful use,
but also should use the discussion of this technique to consider the
future of forensic use of DNA. This bill, which combines a requirement
that the Justice Department facilitate the technique's use with
discretion for the Attorney General to determine the determine the
proper ways to use it, is a good way to proceed.
---------------------------------------------------------------------------
\4\ Henry T. Greely, Daniel P. Riordan, Nanibaa' A. Garrison,
Joanna L. Mountain, Family Ties: The Use of DNA Offender Databases to
Catch Offenders' Kin, Journal of Law, Medicine & Ethics, 34:248-262
(Summer 2006). Much of the analysis in this testimony is drawn from
that article, although my conclusions are not necessarily shared by my
co-authors on that paper.
---------------------------------------------------------------------------
Other particularly useful articles on this topic include Frederick H.
Bieber, Charles H. Brenner & David Lazer, Finding Criminals Through DNA
of Their Relatives, 312 Science 1315-16 (2006); Sonia M. Suter, All in
the Family: Privacy and DNA Familial Searching, 23 Harv. J. Law & Tech.
309 (2010); Erin Murphy, Relative Doubt: Familial Searches of DNA
Databases, 109 Mich. L. Rev. 291 (2010); and Natalie Ram, Fortuity and
Forensic Familial Identification, 63 Stan. L. Rev. 751 (2011). Bieber,
et al., is the other early discussion of the issue. Suter and Murphy
take a more negative view of the technique than I do and are
particularly worth reading; Ram provides some actual data about
different state policies on the method.
Other published legal articles and notes on the topic include Lina
Alexandra Hogan, Note: Fourth Amendment--Guilt by Relation: If Your
Brother Is Convicted of a Crime, You Too May Do Time, 30 W. New Eng. L.
Rev. 543 (2008); Kimberly A. Wah, Note and Comment: A New Investigative
Lead: Familial Searching as an Effective Crime-Fighting Tool, 29
Whittier L. Rev. 909 (2008); Jules Epstein, ``Genetic Surveillance''--
The Bogeyman Response to Familial DNA Investigations, 2009 U. Ill. J.
L. Tech. & Pol'y 141; Jessica D. Gabel, Probable Cause from Probable
Bonds: A Genetic Tattle Tale Based on Familial DNA, 21 Hastings Women's
L.J. 3 (2010); Brett Mares, A Chip Off the Old Block: Familial DNA
Searches and the African American Community, 29 Law & Ineq. 395 (2011);
Amanda Paddock, It's All Relative: Familial DNA Testing and the Fourth
Amendment, 12 Minn. J. L. Sci. & Tech. 851 (2011); Mary McCarthy, Am I
My Brother's Keeper?: Familial DNA Searches in the Twenty-First
Century, 86 Notre Dame L. Rev. 381 (2011); and Jenny Choi, California
and the Future of Partial Match DNA Investigations, 39 Hastings Const.
L.Q. 713 (2012).
I want to do five things in this testimony. First, I will explain
how family forensic DNA works. Second, I will discuss its weaknesses as
a law enforcement tool. Third, I will describe the possible legal and
policy issues this tool raises and how they might (and might not) be
mitigated. Fourth, I will discuss some possible ways to improve the
effectiveness of the technique, though perhaps at the cost of
exacerbating some of its problems. And, finally, I want to reflect on
the trajectory of our use of forensic DNA and where that trajectory may
eventually lead us.
how it works
Each human has two complete human genomes, one inherited from his
or her mother and one from the father. The information in each is
contained in about 3.4 billion ``base pairs''--molecules of adenine
(A), cytosine (C), guanine (G), and thymine (T). Each A is paired with
a T; each C is paired with G. Together, these base pairs form the
``rungs'' of the spiraling staircase that is DNA. Almost all this DNA
is tucked away in the 46 chromosomes in the nuclei of our cells, 22
pairs of ``autosomes,'' cleverly named chromosomes 1 through 22, and
two ``sex'' chromosomes, the X and Y-chromosomes. Men have one X
chromosome, inherited from their mothers (who only have X chromosomes),
and one Y chromosome, inherited from their fathers. Women have one X
chromosome inherited from their mothers and a second X chromosome
inherited from their fathers.
If we think of each base pair as a letter, the ``book'' that is
each of our genomes is about 6.8 billion letters long. This is roughly
the same length at two complete copies of F.2d--not of one volume, but
from the first word of 1 F.2d through the end of 999 F.2d. The copy the
human genome that each of us has is almost entirely identical to the
copy found in any other human--we differ in only about one base pair in
a thousand, so our genomes are roughly 99.9 percent identical. But,
with 6.8 million base pair, that 0.1 percent difference comes out to
about 7 million differences.
Forensic DNA uses those differences to say that crime scene DNA
``matches'' the DNA of a particular suspect. The chances that two
different people (who are not identical twins) would have exactly the
same DNA are infinitesimal. But with 6.8 billion base pairs, where
should we look for differences? In the mid-1990s, the FBI decided to
focus its identification efforts on 13 particular locations in the
genome. These locations, known as ``loci'', are often referred to as
the CODIS loci, because the FBI uses them in its Combined Operating DNA
Information System (CODIS).
The FBI chose thirteen loci where our genome ``stutters.'' These
are short tandem repeats, sometimes called satellite tandem repeats. A
CODIS locus might, for example, consist of a stretch of chromosome 8
where a four base pair sequence, say ATTG, repeats itself. On some
copies of chromosome 8, there might be seven repeats; on others, three
repeats; and on still others, twelve repeats. These thirteen CODIS loci
are all found on the autosomes (chromosomes 1 through 22), so each of
us has two copies of the each of those chromosomes, and so two copies
of each locus--one inherited from our mother and one from our father.
On one locus, for example, I might have five repeats on one chromosome
and eight on another. On another, I might have six repeats on one and
eleven on the other. My CODIS profile is thirteen pairs of numbers, two
for each of the thirteen loci, where each number represents the number
of times a sequence of bases repeats.
Those thirteen pairs of numbers are my ``identity code,'' because
the chances that any human being (other than my identical twin), alive
today or at any time during our species existence, shares the same
thirteen pairs of numbers are very close to zero. Assume, for present
purposes, that each of the thirteen loci has ten different sets of
repeat lengths (called alleles), each of which is found in ten percent
of chromosomes. The chance that, at any locus, I would share both of my
alleles (repeat lengths) with anyone else is about two in one hundred.
Two percent is not a very low probability--but now extend that from one
locus to thirteen loci. Two in one hundred becomes roughly 8,000 in 100
septillion, or about one in 10 sextillion--one in
10,000,000,000,000,000,000,000.
The actual percentages are calculated in a more accurate and
complicated way, but this approach leads to courtroom testimony that
the chances that some DNA came from someone other than the defendant
(or his identical twin) are one in many trillions or even quadrillions.
This is the power of DNA for identification and courts (and police,
prosecutors, and defense counsel) have been using it with confidence
for over 15 years.
The FBI did not have to choose these particular CODIS markers. The
United Kingdom, which has an older and (as a proportion of its
population) bigger database, uses ten loci, only some of which are used
by the FBI. The FBI was looking for loci that were easy to analyze,
using the technology of the mid-1990s, and that had a lot of variation
across all humans. Many other short tandem repeats could have been
used, as well as many other kinds of variation in the genome, but the
CODIS markers work perfectly well for identification. When crime scene
DNA is analyzed for its CODIS markers, the resulting profile can be
compared to the CODIS profiles of suspects, or, through a computerized
search, with the CODIS profiles of the roughly 10 million people whose
profiles are in the FBI's Offender Database. A perfect match means it
is almost certain that the crime scene DNA came from the person with
the same recorded CODIS profile.
The Offender Database contains the CODIS profiles that Congress has
authorized the FBI to collect and include, both from the federal
judicial system and from state systems. The boundaries of the CODIS
system have changed over the years, but they now include profiles from
people whose DNA is authorized by federal or state law to be collected
and put into such a database. These may be people convicted of various
crimes--at this point, all felonies and some misdemeanors--or people
arrested for felonies, or non-U.S. nationals detained under federal
government authority. All the profiles must include the CODIS markers
and states submitting profiles to CODIS have to meet various
requirements. As of February 2012, the Offender Database in the
National DNA Index in CODIS contained over 10,560,300 profiles. The FBI
reported that the database had assisted over 166,700 prosecutions
during its existence. This assistance had been provided when a profile
determined from crime scene DNA had been checked against the CODIS
Offender Database and a match had been found.
But what happens when a match is not found? Is the database then
useless?
Note that in all the above discussion, I have excepted identical
twins. Identical twins have the same genomes and hence the same CODIS
markers. They are a special case of family forensic DNA--if crime scene
DNA matches perfectly the profile of someone in the Offender Database,
but that person could not have been the perpetrator (because, for
example, he was in prison at the time of the crime), but he had an
identical twin, that match could implicate the twin.
Most of us do not have identical twins, but we all have or had
parents and many of us have siblings or children. Our genetic first-
degree relatives--parents, siblings, or children--do not share all of
our genetic variations (unless they are identical twins) but, on
average, they share half of them. Two people randomly chosen from the
population will, on average, share eight to nine of the 26 CODIS
alleles; two first-degree relatives will, on average, share 15 to 17 of
them. This is because relatives get their variations from the same
people. Two genetic brothers must have inherited their CODIS markers
from among their parents' markers. If, for one marker, one parent had
six and eight repeats and the other parent had three and eleven
repeats, the siblings must have either a six or an eight or a three or
an eleven. On average, at any given locus, they will have identical
markers 25 percent of the time, they will share one marker 50 percent
of the time, and they will share neither marker 25 percent of the time.
In fact, because their parents will sometimes have the same
alleles--one parent has, say, five and seven repeats at one CODIS locus
and the other has five and nine--siblings will, on average, share more
than 13 alleles. In the European-American population, siblings will, on
average, share both alleles at five CODIS loci, share one allele at
seven CODIS loci, and share no alleles at one CODIS locus. Thus, on
average, they will share 17 alleles.
The pattern for parent-child matches is a little different. Every
child must have at least one allele from each genetic parent. If one
compares the CODIS profile of a father and son, the son must have one
of the father's alleles at each of the thirteen CODIS loci--because he
got one of his two alleles at each locus from his father. Again,
because the father and mother are likely to share some alleles, the
actual average match between father and son will be more than 13
alleles. Among European-Americans, the average parent and child will
match on about 15.7 alleles. This is fewer than the average siblings,
but the parent-child pattern is distinctive; unlike the siblings, a
parent-child pair must match at at least one of the two alleles at each
locus.
This is the key to family forensic DNA. If crime scene DNA does not
perfectly match the profiles of anyone in the Offender Database, it
might match some of those profiles much more closely than one would
expect. That might be a result of chance--or it might be the result of
the crime scene coming from a close genetic relative of the person in
the Offender Database. The close relatives of the person in the
Offender Database could become leads, to be investigated to see if they
might have been the source of the crime scene DNA. An interview might,
for example, establish if the relative had a solid alibi or not. If
enough evidence were collected to provide probable cause, the
relative's DNA could be taken and directly tested to see if it matched
the crime scene DNA. The partial, family match would no longer be
relevant. The suspect's DNA profile either would or would not match the
crime scene DNA profile; the family match would have only been the
reason to investigate this person, it would not actually be evidence in
court against him.
The British became using family forensic DNA as an investigative
technique nearly a decade ago, with occasional success. At least two
high profile American cases have used variations on family forensic
DNA. The Grim Sleeper case from Los Angeles is the purest example. The
suspect was ultimately identified because the profile of the crime
scene DNA bore a close resemblance to the DNA profile of his son, who
was in the Offender Database as a result of his own run-ins with the
law. The police interviewed the son, learned that his father had lived
in the area of the crimes, and proceeded to investigate and ultimately
arrest the father.
This use of family forensic DNA, the kind most commonly
contemplated, basically asks the CODIS Offender Database, ``are there
people in the database whose DNA profiles indicate they are likely to
be closely related to the person who left the crime scene DNA.'' Unlike
traditional CODIS searches, these will not turn up perfect matches, but
only partial matches, but matches that are sufficiently good to raise
an inference of a family relationship.
The term ``partial match'' needs to be used with care. ``Partial
match'' has meaning in forensic DNA totally apart from family forensic
DNA. In some cases, the crime scene DNA is degraded or damaged and not
all 26 alleles can be derived from it. If, for example, only 20 alleles
can be analyzed from the crime scene DNA and a suspect matches on those
20 alleles, this raises questions about what the odds are that the
match is not coincidence. Family forensic DNA presupposes having all
the alleles from the crime scene DNA but only having some match;
``partial matching'' has usually meant having only some of the alleles
from the crime scene DNA but having all of those alleles matching. I
believe there has been some confusion about whether state regulations
governing partial matches were meant to apply to family forensic DNA.
The BTK case from Kansas provide a somewhat different example of
using family forensic DNA. The police had plentiful crime scene DNA,
but when they finally identified a suspect, they had no DNA from him.
They got a court order to force a health clinic to provide a tissue
sample from the suspect's daughter. This was then checked to see if the
crime scene DNA could have come from her father. When they concluded it
could have, they got a DNA sample from the father, which matched the
crime scene DNA. A guilty plea followed.
the weakness of family forensic dna
The biggest weakness of family forensic DNA is that, as an
investigative technique, it is just not very good. It will almost
always produce many false positives, people whose DNA profiles indicate
that they could have family members who left the crime scene DNA but
who did not. Additionally, the technique can produce false negatives,
by not finding people whose close relatives actually did leave the
crime scene DNA.
The false positive problem is large. Although, on average, parents
and children will share about 15 to 16 alleles and unrelated people
will share about 8.7 alleles, some unrelated people will share more
than 8.7 alleles--some, in fact, will share more than 16 alleles. The
larger the number of profiles in the database, the greater the chance
of false positives.
Consider, for example, father-son matches. Each son must match his
genetic father at at least one allele at each locus. What happens if
one asks the CODIS system to identify everyone in the Offender Database
who could be a parent (or child) of the source of the crime scene DNA--
everyone who has at least one allele identical to the crime scene DNA
at each of the 13 loci? In 2006, we calculated that the chance that
crime scene DNA with an ``average'' set of alleles would be consistent
with a parent-child relationship with a random profile from the
Offender Database. We concluded that a DNA profile of average rarity
would be a ``parent-child'' match to between 2,000 and 3,000 profiles
in the Offender Database. When we made those calculations, the Offender
Database had 2.75 million profiles; today it has over 10.5 million
profiles. The average crime scene DNA should now produce 7 to 12
thousand possible ``relatives.'' All or all but one of them will be
false positives. If the crime scene DNA has a particularly rare set of
alleles, there may be no false positives; if it has the most common set
of alleles, there may now be over 100,000 false positives. And as the
Offender Database gets larger, these problems will only get worse.
Of course, one could cut down on false positives by tightening the
requirement for a match. Instead of just requiring one match at every
locus in order to raise suspicion of a parent/child match, one could
require one match at every locus plus two matches at two or three loci.
This is in line with the average number of matches expected, but it
means that a true family match might be missed. If the parent (or
child) of the actual source of the crime scene DNA is in the Offender
Database, he might match at only 13 or 14 loci, not 15 or 16. The
higher we set the bar, the fewer the false positives, but the greater
the risk of false negatives.
This is even truer of sibling/sibling matches, as siblings do not
have the same kind of minimum match as parents and children do. Two
siblings could, in theory, match at every allele or match at none. If
one set the standard for a possible sibling/sibling match so that it
would have a false positive rate similar to that discussed above for
parent/child matches, about 40 percent of actual sibling matches would
be missed. If one set the false positive rate much lower, at, say, one
in 100,000, leading to less than 100 false positives on average with
today's Offenders Database, one would miss about eighty percent of
actual sibling/sibling matches.
As with most tests, there is an inevitable trade-off. The lower the
rate of false positives, the higher rate of false negatives, and vice
versa. But there is yet another problem with the accuracy of family
forensic DNA. We have been talking about false negatives on the
assumption that the source of crime scene DNA actually has a close
relative in the Offender Database but that the comparison does not
reveal the relationship. If the source of the crime scene DNA does not
have a close relative in the Offender Database, the only positives that
family forensic DNA could find would be false positives.
Family forensic DNA is, therefore, not a very good source of leads.
It will usually throw up a vast number of possible suspects and,
depending on where the line is drawn, it may well miss the actual
perpetrator. It will almost always require substantial traditional
police work to follow up the leads, work that, unlike a family search
on CODIS, will eat up scarce police resources. It may be useful in high
profile and difficult crimes, it may be difficult in crimes where the
crime scene DNA has a particularly set of variations, but it will not,
at least as currently feasible, put a major dent into crime.
issues with family forensic dna
As set out in detail in our 2006 article, there seem to be no
strong constitutional or other legal objections to the use of family
forensic DNA. At first glance, it might seem to run afoul of the broad
legal prohibition of ``corruption of blood,'' both in the Constitution
(for the crime of treason) and in the constitutions and statutes of
many states. But those prohibitions concern punishing innocent people
for the crimes of their relatives, not of making people potential
suspects based the crimes of their relatives.
If, in a line-up, the victim says ``the mugger was not number 3,
but he could have been his brother,'' nothing prevents the police from
investigating to see if ``number 3'' has a brother and his whereabouts
at the time of the crime. Similarly, relatives of organized crime
bosses are likely to be under increased suspicion of involvement in mob
crimes. Family relationships are a clue that may properly lead to
investigation. It feels ``unseemly'' to make someone a suspect based on
the crimes of his relatives, but I see no good argument that it is
unconstitutional, or even, in general, a bad idea.
This conclusion is particularly strong in the DNA context, where a
false positive family connection can almost certainly not lead to a
false conviction. Once a relative is identified, his DNA can be taken
(voluntarily or, with probable cause, by legal action) and compared
with the crime scene DNA. If he did not leave the crime scene DNA, no
matter how closely the crime scene DNA matches that of his relative in
the Offender Index, it cannot match his own DNA. The DNA evidence must
exonerate the false positives.
The chance of false conviction, however, is not the only cost to
being falsely identified as a suspect. Being interviewed by the police
will often be a time-consuming and stressful experience, even for
people who know they are innocent. The family suspect may not seriously
risk false conviction, but neither will he be compensated for the time,
anxiety, and possible embarrassment the investigation causes.
Three other issues deserve mention: the possible revelation of
family secrets, possible unfairness to groups that are relatively
genetically homogenous, and possible unfairness to groups that are
disproportionately represented in the Offender Database.
Family forensic DNA is using possible family relationships to look
for suspects. By looking at genetic evidence for family relationships,
though, the technique could reveal facts about those relationships that
are unwelcome, unknown, or both. These facts are most likely to involve
so-called ``false paternity''--the situation where a child's genetic
father is not, as a result of adoption, sperm donation, or other sexual
partners, the man accepted as the genetic father. (The ``preferred''
term for this is ``misattributed parentage,'' but ``false maternity''
is, for understandable reasons, quite rare.)
It is easy to find geneticists who will say that in various genetic
studies, five to ten percent of children have ``unexpected'' genetic
fathers. There is a real dearth of actual published evidence on the
frequency of false paternity and some of the published evidence points
to much lower rates. It does seem likely, though, that the rate is high
enough to be non-trivial--and disconcerting to men who think they are
genetic fathers.
It is possible that family forensic DNA could reveal cases of false
paternity. If, for example, crime scene DNA is consistent with the
suspect being the son of a particular person in the Offender Database,
the ``offender'' could be asked about his children and his sons could
then be questioned. If the interrogation included a DNA sample, its
analysis might show definitively that there can be no parental link
between the two men. Analysis of the CODIS markers could not rule out,
definitively, the possibility that two people were siblings, but could
make that result extremely unlikely.
This information might, or might not, already be known, or
suspected, by those involved. If the investigators do not reveal it, it
seems no concrete harm would be done, though people could still be
understandably upset that the government learned either secret or
previously unknown about their family connections. Alternatively, if
the family members subsequently discovered the government had this
information, they might complain that they were not told; there are,
for example, some potential medical benefits to having an accurate
understanding of one's family history. If the investigators did reveal
the information, though, the chances of disruptions of the family
ties--and perhaps even of violence directed against the mother--seem
quite real. There seems to be no investigative reason to disclose the
results; if the son or brother presumably was ruled out as a suspect by
the DNA analysis whether or not he was related to the person in the
Offender Database. Prohibiting, or greatly limiting, the dissemination
of this kind of family relationship information seems proper.
Second, some populations are more closely genetically homogenous
than others. A small and relatively isolated Native American tribe or a
group of immigrants from one community in, say, Southeast Asia, for
example, is likely to have much closer family relationships, and hence
much more genetic similarity, than, say, ``European-Americans'' or even
``Irish Americans.'' If the crime scene DNA came from a member of such
a population, a higher percentage of people from that group who are in
the Offender Database will be indicated as possibly related to the
source of the crime scene DNA. Law enforcement should be aware of this
bias in the method and treat the community sensitively.
Finally, and, to my mind, most importantly, the results of family
forensic DNA searches of the CODIS Offender Database will be skewed in
the same way that database is skewed. Most notably, African-Americans
are convicted of felonies at roughly three times the rate of their
roughly 13 percent share of the population. One can debate endlessly
the reasons for this disproportion; for present purposes it is enough
that it exists. The result is that, on average, a higher percentage of
the African-American population is likely to be closely related to
someone in the Offender Database than of most other American
populations. The people identified as potential suspects by this method
are therefore much likelier to be African-American than people randomly
chosen from the population. This could be seen as unfair ``special
surveillance'' of the African-American population, and particularly of
innocent members of the population whose only suspicious action is to
share DNA variations with someone in the Offender Database.
At the same time, African-Americans are already likely to be
suspects at a disproportionate rate, for whatever reasons lie behind
the conviction disproportion. And much of the crime committed by
African-Americans victimizes other African-Americans. Still, widespread
use of family forensic DNA, with its vast number of false positives
bringing under suspicion many innocent people, could well be seen by
many African-Americans as another ``racist'' action by the American
criminal justice system. Although these concerns about family forensic
DNA do not seem to me to rise to the level of a possible constitutional
violation, the public reaction could still be real and problematic.
possible improvements in the effectiveness of family forensic dna
The biggest problems of family forensic DNA stem from its
inaccuracy. It is likely to be throw up so many possible family
connections that its use will often impose costs, in police time and in
the costs to innocent family members of being, even briefly, suspects,
as to limit its use to only very unusual cases. These would be cases
where the rarity of some of the alleles in the crime scene DNA greatly
limits the number of ``hits'' or where the difficulty and importance of
solving the crime justifies spending great resources. This inaccuracy
can be combated, in ways both mundane and scientific, though these
solutions raise their own problems.
One problem in implementing family forensic DNA is the need to find
out whether someone in the Offender Database who is identified as a
possible relative of the source of the crime scene DNA in fact has any
relatives who might have been that source. This will typically involve
finding and interviewing the ``offender,'' as well as hoping for his
cooperation. This step could be eased if a computerized record existed
of the relatives of everyone in the Offender Database. A simple
questionnaire at time of conviction or arrest could provide such
information and then a database search could quickly narrow down the
possible family connections to only those with relatives--and could
give priority to investigating the families of those who have relatives
of the expected sex, age, and geographic location to have been involved
in the crime. The problem is that it seems hard to justify asking a
newly arrested or convicted person about his relatives and even harder
to make a case that answering such questions should be required.
The technical approaches are more promising, but they, too, have
problems. CODIS is just not a very good system for determining family
relationships. With only 26 alleles, the chances are fairly good that
some non-relatives will randomly match the crime scene DNA on enough
alleles to signal a possible family relationship. That chance grows
with the Offender Database. This is the fundamental cause of the vast
number of false positives with this technique.
Using more alleles can make the process much more accurate. Our
2006 paper calculated that by adding 20 more loci similar to the
existing CODIS markers, the chances of a false parent/child match would
be about one in 200 million, reducing the number of false positives
from hundreds or thousands to a handful or fewer.
California's implementation of family forensic DNA uses a similar
expansion of alleles to narrow the number of false positives. It
requires the authorities to check Y chromosome markers from the
offender and the crime scene DNA and only authorizes proceeding to
investigate the family match if the Y chromosome markers also match.
Men inherit their Y chromosomes from their fathers. If two people have
identical sets of markers on their Y chromosomes, they are very likely
to share an ancestor in their paternal line. They might be father and
son, brother and brother, or cousins who are both the sons of brothers.
They may also be more distantly related, but the Y chromosome is
sufficiently variable in human populations that exact Y chromosome
matches will be rare. The existence of a Y chromosome match does not
itself indicate guilt--innocent brothers will share the same Y
chromosome--but use of Y chromosome matching will pare down the number
of leads enormously, again reducing the number of false positives. This
both improves the efficiency of the process for the police and cuts the
number of innocent people who will be, however briefly, suspects.
The alleles examined on the Y chromosome share with the CODIS loci
the virtue of having no known (or likely) medical or physical
consequences. They seem to be so-called ``junk'' DNA, useful only for
identification. One problem with the Y chromosome is that it is only
found in men. Neither crime scene DNA from a woman nor the DNA of any
women in the Offender Database could be checked against the Y
chromosome. As over 90 percent of convicted felons are male, this is a
concern, but not a huge one. And other parts of the genome that are
similarly variable to the Y chromosome could be checked from women.
A bigger problem with using Y chromosome matching as part of family
forensic DNA, though, is that the Y chromosome alleles have not been
analyzed for the 10.5 million people already in the Offender Database.
To do that analysis would require either re-analyzing the saved DNA
sample the ``offender'' earlier provided--if it was saved--or acquiring
a new sample. The costs of doing that for over ten million people, or
even of finding many of them, would be quite high. On the other hand,
one could do it a case at a time, seeking to analyze only the Y
chromosomes of those ``offenders'' picked out by the family forensic
analysis. This requires DNA from those ``offenders'' to be readily
available or to be easy to re-acquire. It is hard to see a
justification for forcing an offender to provide another DNA sample to
investigate a crime that, as the result of the lack of an exact match,
we know he cannot have committed. It might be possible to obtain a
search warrant requiring a new DNA sample, but the more positive family
matches there are, and, as a result, the lower the chance that any one
of the offenders involved in those positive matches is actually related
to the source of the crime scene DNA, the harder the case would seem
for show probable cause.
One could also use other technical solutions. A common tool for
genetics and genomics research, with some commercial uses, is the so-
called ``SNP chip.'' This device allows the operator to determine,
cheaply and quickly, which base (A, C, G, or T) a person carries at
locations known as ``single nucleotide polymorphisms'' (``SNPs''),
where substantial percentages of the population carry different bases.
These SNP chips can quite easily examine hundreds of thousands or even
millions of these SNPs. While the chance that two unrelated people
might share 13 of the 26 CODIS alleles by chance is not necessarily
small, the chance that two unrelated people would share 300,000 out of
600,000 SNPs is vanishingly small. SNP chips could determine the
existence of a wide range of relationship, not just first-degree
relationships like parent/child or sib/sib, but uncle/nephew, cousin/
cousin, and others. SNP chips could easily replace the CODIS loci
entirely.
This solution, though, also has problems. It shares one with the Y
chromosome tactic--it would require re-analyzing DNA from the entire
10.5 million person Offender Database in order to use it to search that
database. But it has another problem. Unlike the CODIS loci or the
commonly analyzed Y chromosome markers, many of these SNPs are
associated with particular diseases or other genetic traits. Doing a
SNP analysis for forensic purposes does raise all the privacy questions
that are avoided when the genetic variations being used seem to be
useful only for identification.
Improving the efficiency of family forensic DNA is both possible
and, if the method is to be used at all, valuable both to police and to
innocent potential suspects. If this bill passes, the Attorney General,
in promulgating regulations, should give serious consideration to these
ways to minimize false positives. But each of them poses serious
challenges.
the trajectory of forensic dna
I cannot leave this topic without noting the trajectory of forensic
DNA use. Governments initially required DNA samples from people
convicted of the most serious felonies, usually murder and sexual
assaults. Then they began to require DNA samples from people convicted
of less serious felonies or of serious misdemeanors or from juveniles
found delinquent for reasons that would, had they been adults, been
felonies or serious misdemeanors. More recently, first states and then
the federal government required DNA samples from people charged with
felonies, whether or not they were then, or ever, convicted. (The
constitutionality of these statutes under the Fourth Amendment
continues to be debated in federal and state courts across the
country.) Federal legislation now authorizes the mandatory collection
of DNA from non-U.S. persons ``detained'' under the government's
authority, whether or not charged with a felony or any crime. And just
last month, the State of New York passed legislation requiring DNA
samples from people convicted of most misdemeanors.
The trajectory has clearly been to collect more and more DNA from
people with decreasingly serious involvement with the criminal justice
system. Advocates have argued, and most judges have agreed, that people
with those connections to the criminal justice system have forfeited
some of their rights as a result of their convictions, arrests, or
detainments. Family forensic DNA is a technique that uses the DNA
provided under those statutes to extend the reach of forensic DNA to
people who have not necessarily had any contact with the criminal
justice system, let alone conviction or charge--people whose only link
is that they are related to people who were convicted or arrested or
detained. It is a logical and scientifically useful outgrowth of the
earlier collections, not, I think, a planned consequence of those
databases but a clever way to use them to solve more crimes, based on
the reality that genetic variations run in families. This bill would
take that informal and almost accidental growth and give it the force
of law, providing a legislative endorsement of the extension of
forensic DNA to catch people who had no prior record of conviction,
arrest, or detention.
This makes sense as a way to catch more criminals and its costs to
the innocent are low. But if we really want to maximize the value of
forensic DNA, why stop with (the usually innocent) first degree
relatives of those convicted, arrested, or detained? The logical size
for a forensic DNA database, at least once forensic DNA is cut loose
from its mooring to an individual's involvement in the criminal justice
system, is universal. A truly universal forensic DNA database would
make family searching obsolete--the family members you might find would
already be in the database. And it would also end the ways family
searching discriminates against people whose family members were
convicted, arrested, or detained.
In fact, an unplanned and impromptu version of such a universal
database may be on its way. The cost of genomic analysis, and even of
sequencing a person's entire genome, has been falling dramatically. The
medical value of that information has been increasing steadily,
although, unfortunately, not as dramatically. Within a decade scores,
if not hundreds, of millions of Americans will have substantial genomic
information in their clinical electronic health records, information
that will be perfectly useful for identification--and that is only a
court order away from the scrutiny of the government (or, in some
cases, private litigants).
A universal DNA forensic database seems to me politically
impossible today. At any time, such a database would be fraught with
concerns about privacy and misuse. (I would note that restricting such
a database to genetic information useful only for identification and
not for any other purpose, unlike the information in medical records,
would be a useful way to handle some of those concerns.) Whether such a
database could be justified as a matter of policy would depend
crucially on the protections that came with it. Whether a mandatory
universal database could be justified constitutionally is another
thorny question; my guess is that it would not be upheld as a mandate
but might be upheld as a condition to participation in some
governmental program for which definite identification is useful, like
a driver's license, Social Security, or Medicare.
Foreign countries and political leaders have toyed with the idea of
a universal DNA database, including the democratic government of
Portugal and the former prime minister of the United Kingdom, Tony
Blair. It is not a question for this subcommittee today, or, I suspect,
any day soon. But endorsing the use of family forensic DNA and using
DNA to make suspects of people with no prior personal connection to the
criminal justice system takes us one logical step toward a universal
DNA database--and would make the day when that discussion is necessary
draw nearer.
conclusion
I support H.R. 3361, although with reservations. It is not a
panacea. It will not solve a large number of crimes, but it will solve
some crimes, at some cost to the public in convenience, in privacy, and
in their presumed innocence. If managed well by the Attorney General's
regulations, that (small) cost seems to me likely to be a cost that is
likely to be outweighed by the technique's (also fairly small)
benefits. The technique should be used responsibly and ways to improve
it--for the benefit of both the police and the public--should be
explored and debated. But this bill should also make us think about
where we want the use of forensic DNA techniques and databases to go.
If all the bill does is to spark a realistic discussion of that
question, that alone may make it worthwhile.
__________
Mr. Sensenbrenner. Mr. Risher.
TESTIMONY OF MICHAEL T. RISHER, STAFF ATTORNEY,
AMERICAN CIVIL LIBERTIES UNION FOUNDATION OF NORTHERN
CALIFORNIA
Mr. Risher. Thank you, Mr. Chairman. The virtue of criminal
DNA evidence in our criminal justice system is that it is
incredibly precise as we just heard from Professor Greely. The
problem with familial searching is it turns that precision on
its head, it is inherently imprecise. We are not finding the
exact match in the system, instead we are finding people who
are kind of close, they may be family members. And it is this
inherent imprecision that leads to many problems, regarding
privacy, efficiency, efficacy and basic fairness.
We should remember that we are talking here not about
whether familial searching should be used in individual cases
by the States, but whether the FBI should be using it
nationally. And using it nationally generates particular
problems. So what are some of those problems?
The imprecision means that familial searching can invade
privacy, familial privacy, personal privacy, genetic privacy.
Law enforcement may find out information about blood
relationships that family members are completely unaware of.
The families may learn of those; that can be incredibly
disruptive to families. It can lead to investigation,
interrogation, even harassment of perfectly innocent family
members simply because they have a family member who has been
arrested or convicted.
The imprecision means the familial searching has monetary
costs. It costs money to go and retest 800, 1,000 samples which
is a necessary part of familial searching. We don't have an
infinite amount of money for our criminal justice system, and
that money could be better spent. We could test some of those
thousands upon thousands of rape kits with money that otherwise
would be used for familial searching.
The imprecision means that familial searching is inherently
unfair. We have never, in our country, had a system that
focuses suspicion on people because of their blood, this is
guilt by blood as some people have called it. It is going to
affect people with larger families, it is going to affect
people of color, it is going to affect poorer people, because
those groups are more likely to have a relative who is in the
CODIS system either because of an arrest which may never have
been prosecuted or because of a conviction.
And it is simply unfair to focus on one subgroup of people
for those reasons, while potentially ignoring people who have
done the same or worse crimes who are fortunate enough not to
have a relative in the CODIS system.
Finally the imprecision means that familial searching is
not particularly effective, and unlike--we heard, I think, one
case out of 13 in California where I practice where it came up
with results, we don't know, we don't have the information to
see how much cost in terms of money, societal costs were
incurred in those other 12, unsuccessful investigations. The
odd thing about familial searching, as you have heard, is with
most DNA databanks, to some degree, the larger the databank,
the more benefit you get. With familial searching, that is not
true. If you have an enormous databank, you get so many false
positives, so many false near hits, that the process becomes
less efficient, which is an important consideration when
talking about expanding it to the national level as opposed to
leaving it to the States as is done right now.
So because of this imprecision, and because of these
societal and monetary costs, Congress should consider the issue
very closely before expanding familial searching, taking it
away from the States in a sense, and giving to the FBI to do
and the national databank. Quite frankly, with the available
information which is not great, we take the position that the
costs outweigh the benefits of taking this national.
If Congress does disagree and decide to authorize the FBI
to use this technique it needs to ensure that civil liberties,
that familial privacy, and that efficacy are protected. And the
draft bill does that in some ways. It requires that it be used
only for very serious cases. It requires the States certify
they have privacy protections in place; it requires reporting
on the results. But there are two big gaps. And one of them is
other investigative techniques be used and exhausted first.
California does that, I hear Virginia does that, that is in the
Wiretap Act, that should be added to any bill under
consideration.
And finally we need judicial authorization. The history of
this country, our Constitution showed that when privacy is
going to be invaded in a serious way, the decision to do that
should be made by a neutral magistrate, not by law enforcement
officers who, of course, have a great incentive to try to solve
crimes, that is their job; it is the job of a judge to weigh
the value of crime solving techniques against the cost to
society.
So that is where we come out. If Congress does decide to
authorize this technique, I think it should very seriously
consider the cost and benefits and should protect privacy in
the statute rather than leaving it up to the regulations, thank
you.
[The prepared statement of Mr. Risher follows:]
Prepared Statement of Michael T. Risher, Staff Attorney,
American Civil Liberties Union Foundation of Northern California
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
__________
Mr. Sensenbrenner. I want to thank all of the witnesses for
their testimony, which has been very enlightening, both in
terms of the technical side and the legal side. I am going to
recognize myself for 5 minutes and hope that I don't have to
use all of it. When I was an undergraduate at Stanford, I got
quite an interest in Russian history with all the archives that
have been in the Hoover Institution. And after I left and got
involved in this business, the bones of Czar Nicholas the II
were supposedly identified in part with a DNA match with the
Duke of Edinburgh, Prince Philip, who was a fifth or sixth
cousin several times removed from the late czar. How accurate
would that be, because this is a case of familial matching, but
with a really very, very distant relative. Professor Greely or
Mr. Marone, you are probably the two best to answer that.
Mr. Greely. I actually know a fair amount about that
particular case, and it was a kind of matching different from
the CODIS matching. What they were looking for there was so-
called Mitochondrial DNA. This is DNA that is not in the
nucleus of our cells, not in the 46 chromosomes, but it is in
little things called Mitochondria, which you may remember from
high school biology as the energy powerhouses of the cell and
they have their own DNA. It is a very small bit but it is
distinctive. And what is really interesting about it is you
only get from your mother. Now if Prince Philip's, great,
great, mother's mother's mother's mother was Queen, I think it
was Queen Victoria.
Mr. Sensenbrenner. Yeah, she was the grandmother of all of
the European royal houses.
Mr. Greely. Which is why they had hemophilia as well. The
same was true with the czar. So since they both got the
Mitochondria of Queen Victoria, they were able to make a match
between the czar's Mitochondrial DNA and Prince Philip's
Mitochondrial DNA. That particular case is especially odd,
though, because unusually, the czar's bones had two different
DNA signatures for the Mitochondria that sometimes happens, it
is unusual. They then got the czar's brother's body out of the
Cathedral of Saint Peter and Saint Paul in Saint Petersburg and
checked, and the brother had the same mix. So they concluded
that yes, in fact, that was the czar's remains.
Mr. Sensenbrenner. So in terms of what is being proposed in
this bill----
Mr. Greely. This is different.
Mr. Sensenbrenner. It is different. I am enlightened. The
gentleman from Virginia, Mr. Scott.
Mr. Scott. Thank you. Detective Kilcoyne, your testimony
says the second search of the database produced a match of the
son of Lonnie David Franklin, you mean close match, not a
match, right?
Mr. Kilcoyne. There was a match.
Mr. Scott. A familial match which isn't a direct match.
Mr. Kilcoyne. Correct.
Mr. Scott. And you were given one name?
Mr. Kilcoyne. No.
Mr. Scott. How many names were you given?
Mr. Kilcoyne. There were two names that were given to us
that were--one was the genetic match which was the son; another
is the Bureau of Investigation had developed information on
another man named Franklin that lived in another county. And
there was no genetic match there, but they had developed that
from research and public records. It was not from the
scientific search.
Mr. Scott. So the scientific gave you one name.
Mr. Kilcoyne. Right.
Mr. Scott. Just the son. How many other relatives could
that have implicated?
Mr. Kilcoyne. Well, Mr. Franklin--I don't really know, I
don't know if he has got brothers. He does not have another son
that I am aware of. And that was not our mission to go off on a
genetic witch hunt of the Franklin family.
Mr. Scott. Mr. Marone, when you get a search and get it
down to one person, what is the confidence level that that is
actually a relative?
Mr. Marone. Let me back up and explain how you get down to
one name. What do you with the additional testing with the Y
chromosomes, you are eliminating all of folks are who aren't
related, you may get a couple of brothers, but you are getting
it down to that name. The confidence on that is pretty good
because then you go back, and again, that is still
investigative. You still have to get a sample from the
individual, in their case it was the pizza crust for the
individual.
Mr. Scott. But----
Mr. Marone. You can identify that. That case----
Mr. Scott. If you get down and say this is a relative of
the perpetrator, how confident are you that it is, in fact, a
relative of the perpetrator and not a false positive?
Mr. Marone. I would not--I would not feel comfortable
giving you a percentage number on that.
Mr. Scott. High confidence level, but you can't give a
percentage--sorry. Mr. Greely.
Mr. Greely. If I may, it depends, that is the right start
to any answer from a lawyer. But if you are using the Y
chromosome match in addition, which you can't get directly from
the CODIS database because the Y chromosomes ample types are
not in the CODIS database. If you are using the Y chromosome
match, you can be relatively confident. I would say--I am not
going to give you a percentage either, but it is a pretty high
degree of confidence, but it is not a perfect degree of
confidence. In the long you can be confident they are related,
because all of us are related, all of us are cousins.
People with the same Y chromosomes will be a little more
related. Will they actually know they are related or not?
Sometimes yes, sometimes no. If you are doing the search
without the Y chromosome, the vast majority of people you will
throw up will not be relatives. If you then narrow it down with
the Y chromosome, a very large, perhaps majority, certainly
large minority of the people you discover will be relatives.
Mr. Scott. Mr. Risher, are there any unreasonable search
and seizure implications with this technique?
Mr. Risher. There absolutely are. And perhaps the most
obvious one is that this retesting of the Y chromosome that you
have been hearing about that is a necessary part of this
technique is under, I think, pretty well-established law a new
search. And the justification for taking DNA from a convicted
person or an arrestee may no longer be valid to conduct a new
search on that person's genetic profile that has been sitting
in storage 5, 10, 15 years later. That really can raise pretty
grave concerns when we are talking about retesting genetic
information that was seized from somebody with one
justification when that justification is no longer valid. This
is not an issue that has been litigated, but quite frankly, the
courts may say that that is unconstitutional.
Mr. Scott. Thank you, Mr. Chairman.
Mr. Sensenbrenner. Thank you. Gentleman from Georgia, Mr.
Johnson.
Mr. Johnson. Thank you, Mr. Chairman. Legislation before
us, H.R. 3361, makes no provisions for how familial DNA
evidence would be collected; does it Detective Kilcoyne?
Mr. Kilcoyne. Sir, currently in the State of California,
DNA is collected a number of ways from people.
Mr. Johnson. No, my question is the legislation itself does
not provide for specific means of collecting familial DNA; is
that correct?
Mr. Kilcoyne. Correct.
Mr. Johnson. Okay. And DNA, familial DNA is--how would you
obtain it in addition to what may be in a criminal database of
a State?
Mr. Kilcoyne. Your DNA is collected upon arrest,
conviction, whatever, just like your fingerprint or photograph,
in California. That sample would need to be broken down.
Mr. Johnson. I guess what I am getting at is are there ways
of collecting familial DNA other than from criminal databases?
Mr. Kilcoyne. Not for law enforcement, no.
Mr. Johnson. Now Mr. Risher, people who are walking along
the street can be stopped and questioned by the police without
any authority to do so; is that correct?
Mr. Risher. Well, the police need reasonable suspicion that
a person walking down the street has committed a crime or is
going to commit a crime in order to stop him.
Mr. Johnson. If you just go up and knock on somebody's door
and say, how are you today, do you know Joe Blow? The police
can do that currently without a warrant or without any
reasonable suspicion and without having to warn the person that
they have a right to not answer the question or you have a
right to just walk off without responding.
Mr. Risher. That is correct.
Mr. Johnson. Now, with a DNA sample of a family member, or
with a DNA sample that law enforcement would like to test for
familial DNA to develop clues which could solve a criminal
offense, what is the difference between--or one does not have a
right to exclude familial DNA from being searched, simply
because it is in a police database. If it is in the database,
then it can be searched, or it can be used for a DNA search
without a warrant or anything like that.
Mr. Risher. Well, it is necessary to distinguish between
two types of databases. We have a database that comprises the
genetic profiles that have been developed from testing samples,
and we also have the physical samples themselves that are
stored in a separate database. I think it is probably correct
that testing those computerized profiles is not a separate
search. To the contrary, testing those physical samples, again,
is a search and the 4th amendment does place limitations on the
authority of the police to do that without a warrant.
It may well be illegal to test those samples if the person
is no longer in prison or on parole, the person who provided
that sample. But to get back to your earlier question, there
are really two differences between the standard knock and talk
procedure and what we are looking at here. One of them is that
unlike in that procedure we are focusing on people here because
of some sort of guilt by blood and I think that makes a lot of
us nervous.
Mr. Johnson. Doesn't law enforcement have--law enforcement
is not encumbered from searching criminal databases for DNA
evidence to locate the perpetrator, why should they be limited
in terms of collecting familial DNA evidence, which could lead
to clues which would solve a crime?
Mr. Risher. Well, in part, because the process is
different. You do need to do this retesting which raises
different constitutional issues.
Mr. Johnson. There is----
Mr. Sensenbrenner. The gentleman's time has expired. The
gentlewoman from California, Ms. Chu.
Ms. Chu. Detective, the legislation being discussed today
would authorize the FBI to conduct these familial searches when
investigating crimes that involve murder, manslaughter, and sex
offenses against a minor. As a California law enforcement
official, you know that California has a similar law, but
nonetheless, it differs in two distinctive ways: One, the crime
being investigated must be one of violence and pose critical
public safety implications. And secondly, all reasonable and
viable investigative leads have to be exhausted. What impact do
these additional safeguards have on the effectiveness of using
familial searches as an investigative tool, and how do these
safeguards ensure that these tools are used efficiently?
Mr. Kilcoyne. I think that the systems in place, the
protocols in place that have been placed on us by the Attorney
General of the State of California have significant protections
in place so we are not straying off on this new science that is
in front of us. The Federal bill that has been proposed here, I
would suggest that it has added also violent sexual crimes
against women, not just minors, that should be included. I
think that what needs to be understood is that when there is a
match--and this should not be looked at any different than
rerunning a second search of the fingerprint databank to try to
find a suspect. We are looking at a genetic fingerprint for the
most part.
We are trying to solve the unsolvable, as in the case of
the Grim Sleeper that went on for 25 years. There are close to
18 families that have been affected, that would basically be
the top row of the Chamber here, that if every one of the
chairs up there had lost a daughter to this man.
This is a significant violation of our being, as human
beings. The genetic search is color-blind. It doesn't tell us
what color this person is that we are looking for. There are
some schools that say, well, there is something that we can
tell what color hair, what color eyes the guy has. That has not
been perfected yet. But the bottom line is, when the
information is passed to the law enforcement agency, it is an
investigative lead, period. It is not conclusive that is your
man. It is no different than if someone called--we run a
composite photograph or a clip from a liquor store surveillance
video during a robbery, that the guy down at the corner gas
station resembles your composite or resembles the newspaper
photograph this morning. That needs to be vetted out by
investigation; and that is what is done, such as the case with
the Grim Sleeper. Once we have established probable cause to
detain the man, then the courts get involved, and they
authorize the search of this man's one-to-one genetic profile
against the crime scene evidence.
Ms. Chu. If I could ask something else.
So it sounds like those types of restrictions are something
that is supportable and works in California.
Mr. Kilcoyne. Very much so.
Ms. Chu. And there is another protocol that California has
which is that the analysis and review of DNA takes place
without the knowledge or communication with the requesting
agency or prosecutor. So there is a separation there. And what
is the relevance behind your statement about that? What is the
benefit of separating the scientists responsible for analyzing
the matches from the local law enforcement that is
investigating the case?
Mr. Kilcoyne. I think it is just a precautionary measure
that puts up a barrier between--you know, just like here in
Washington where you have different houses doing different jobs
and probably not a lot of sharing of information all the time.
But it is a protective umbrella that has the State researching
information; first in the laboratory and then using public
records and their investigative arm and then multiple exchanges
back and forth within that State bubble before and if and when
the information is decided that it merits being passed to the
local agency, the law enforcement agency. And I think that is
important. Because this is so new, there needs to be insurance
that, you know, people's rights aren't being violated and there
is no one stronger than myself or the Chief in Los Angeles that
thinks about, you know, we are going to investigate Johnny for
what his daddy did. And that is not what we are after. It needs
to be remembered, all it is is another investigative tool, an
investigative lead for law enforcement to continue the
furtherance of the investigation.
Mr. Sensenbrenner. The gentlewoman's time has expired. The
gentleman from California, Mr. Lungren.
Mr. Lungren. Thank you very much. I am sorry I was delayed
and didn't hear all the testimony but I have tried to go
through the written testimony.
It is interesting that we have these questions before us. I
recall when I was attorney general of California when we had
the very first cold case as a result of a DNA match on a cold
case of a multiple murder in California and found the guy
sitting in--I think it was an Oklahoma City jail for violent
crimes there.
And I also remember representing the State of California
when we had a dispute over an estate left by a rather wealthy
individual who had received treatment from a University of
California medical school, and as a result, left his fortune to
them. And this fellow had spent a lot of time outside the
country. And I think it was the Mariana Islands or somewhere
around there. And there was this suggestion that he might have
fathered some children there. And they questioned whether the
children had a right to the estate. And, of course, he had died
as a result--presumably of a plane accident. We couldn't find
him; couldn't do a DNA match on him versus one of these
children claiming to be his issue. And we finally convinced a
brother to give DNA at the insistence of the other parties to
the action. And it was sufficient for us to reach a settlement
with that young man and several other young men in that area
because we thought there was sufficient evidence that they had
a claim. And the State of California gave up a portion of that
fortune that was going to the medical schools.
So now to see familial DNA used in the criminal context is
most interesting. So Mr. Risher, I guess I would like to ask
you this question: I reviewed your testimony. It appears to me
that you are not totally objecting to the use of familial DNA
in appropriate investigations, but you want to see it couched
in protective terms, such that we avoid some of the problems
that have been articulated here. Is that correct?
Mr. Risher. It is difficult to say. But for example, the
Grim Sleeper case was not an inappropriate use of familial DNA.
Although I do understand--and maybe this is wrong--that the
person arrested does have prior felony convictions. So under
today's regiment, his DNA would be in the database. There would
be no need for familial searching.
The question of whether the costs outweigh the benefits is
ultimately for the legislature.
Mr. Lungren. Correct. But give me the specific privacy
concerns that you have that you think have to be addressed even
though you may not have determined exactly how they should be
addressed.
Mr. Risher. The specific privacy concerns are that, this is
determining blood relationships, and blood relationships do not
always track our family relationships in our society. We might
have adoptions. We might have infidelity, for example. And the
process--the follow-up investigation--that can, first of all,
make this whole familial DNA searching much less effective
because it doesn't necessarily account for infidelity that can
throw a fork into the system. But it can be very damaging. One
paper says it can destroy families if, in the context of a law
enforcement investigation, we have family members who are
suddenly made aware that the person they thought was--persons
they thought were their children are not, in fact, genetically
related to them. In some cultures, that, of course, can result
in horrible domestic violence consequences.
Mr. Lungren. Does that give rise to a constitutional
privacy issue? Or is it your suggestion that it is beyond that
but nonetheless we ought to protect that in view of the
implications that you have just described?
Mr. Risher. Exactly the latter. Congress has repeatedly
created protections for Americans' privacy that are not found
in the Fourth Amendment, and that is why it is necessary for
Congress to act.
Mr. Lungren. Right. And that is what I am trying to find
out. Do you find this a particular assault? Or does it do
damage to the essence of the Fourth Amendment? Or are we
talking about additional protections that you think are
appropriate but that are not mandated necessarily by the
protections in the Fourth Amendment?
Mr. Risher. There are grave constitutional questions raised
by some aspects of familial searching, particularly the
reanalysis of genetic samples that might have been given long
ago. Putting those aside, which of course will have to be
resolved in individual cases--putting those aside, there are
still societal costs that have nothing to do with the Fourth
Amendment that might have something to do with equal protection
or due process that Congress should act to eliminate any
question.
It is better to act proactively and avoid those questions
coming up, that could void convictions or cost enormous
expenses. Congress should do it now so we don't have to address
those questions in the courts because they will come up in the
courts.
Mr. Sensenbrenner. The gentleman's time has expired. I
would like to thank all of the witnesses for their extremely
useful testimony today. This has been a very enlightening
hearing. I think, as Chairman, this bill needs quite a bit of
fixing up before we send it out of Committee, both in terms of
policy as well as in terms of privacy protections. And your
testimony has helped us to be able to work through that maze.
So again, without objection, the hearing is adjourned.
[Whereupon, at 11:10 a.m., the Subcommittee was adjourned.]