DNA IDENTIFICATION: APPLICATIONS AND ISSUES
CRS Report for Congress
DNA Identification: Applications and Issues
Updated January 12, 2001
Eric A. Fischer
Senior Specialist in Science and Technology
Resources, Science, and Industry Division
Congressional Research Service The Library of Congress
This report provides an overview of how the genetic information contained in DNA is used
for identification, and a discussion of issues associated with those uses. It begins by
discussing the unique properties of genetic information that make it a powerful tool for
identification and what is involved in making identifications from DNA. Next is a description
of current federal programs and activities, followed by discussion of issues raised by the
development of this new technology. Major issues include the use of DNA identification in
the criminal justice system (including sample backlogs, databases, and postconviction DNA
analysis), impacts of technological improvements, and privacy. Legislative activity in theth
Report RL30694, DNA Evidence: Legislative Initiatives in the 106 Congress. For
discussion of genetic privacy and discrimination, see CRS Issue Brief IB98002, Medical
Records Confidentiality, and CRS Report RL30006, Genetic Information: Legal Issues
Relating to Discrimination and Privacy. This report will be updated at least annually.
DNA Identification: Applications and Issues
DNA technology can provide useful identifying information in many situations,
such as in solving crimes, determining paternity, and identifying human remains.
Research is resulting in improvements in sensitivity and power and reductions in cost.
Such use and improvements are raising several policy issues.
The use of DNA in identification results from its unique characteristics: It is a
complex molecule, containing much information. Each person has billions of identical
copies. The structure of the molecule varies from person to person and is inherited,
so the DNA of relatives is more similar than that of unrelated people. Also, DNA is
easily preserved with the structure intact.
Identification requires comparing DNA whose source has not been determined
with DNA whose source is known. The first step is to characterize corresponding
DNA sequences from samples. The resulting profiles are then compared. If they
differ, the samples did not have the same origin. If they match, then they could have
come from the known source, or from someone else who has an identical profile. The
science of population genetics provides ways of estimating quantitatively the chances
that the matched DNA could have come from another source.
Databases or indexes are often used in DNA identification. They might contain
profiles of persons whose identity is known, such as convicted felons, or whose
identity is not known, such as from crime scene samples or unidentified remains. The
Combined DNA Index System (CODIS), administered by the FBI, contains both
kinds. When a profile is obtained from a relevant sample, the database can be
searched to determine if a match is found. Thus, a suspect may be identified when a
profile from a crime-scene sample is searched against profiles of convicted felons.
Congress has enacted several laws relating to DNA evidence. The DNA
Identification Act of 1994 (P.L. 103-322) authorized CODIS and a grants program
for state and local laboratories, and addressed quality control and privacy issues. The
Antiterrorism and Effective Death Penalty Act of 1996 (P.L. 104-132) expanded
CODIS and established a grants program that required states, to be eligible, to collect
DNA samples from persons convicted of felony sex crimes. The Crime Identification
Technology Act of 1998 (P.L. 105-521) established a grants program that funds a
broad range of activities, including several related to DNA typing. The National
Institute of Justice (NIJ) and the Bureau of Justice Assistance (BJA) administer those
and other relevant grants programs. The National Institute of Standards and
Technology (NIST), the Armed Forces Institute of Pathology (AFIP), and the Army
Criminal Investigation Laboratory (USACIL) also have significant DNA identification
Policy issues raised by the use of DNA in identification include how best to
eliminate the large backlog of samples awaiting processing for CODIS, whether to
broaden the offenses that qualify, how to respond to the increasing number of
requests for postconviction DNA analysis, how to address privacy issues, and what
impacts the broadening applications of the technology may have.
Introduction ................................................... 1
Why DNA Can Be Used in Identification .............................2
How DNA Is Used in Identification..................................4
Characterization ............................................. 4
Comparison ................................................ 6
Calculation ................................................. 6
Interpretation ............................................... 7
Databases ..................................................... 9
Federal Agency Programs and Activities.............................10
Federal Bureau of Investigation................................10
DNA Advisory Board....................................11
National Institute of Justice...................................11
Bureau of Justice Assistance..................................13
National Institute of Standards and Technology....................13
Department of Defense......................................14
Relevant Public Laws...........................................15
Current and Emerging Issues .....................................16
Law Enforcement and Criminal Justice...........................17
Databases ............................................. 18
Future Directions of the Technology............................21
Relationship of DNA Identification to Medical Genetic Testing .......23
Privacy and Discrimination....................................24
List of Tables
Appropriations for State and Local DNA Laboratory Support, FY1996–FY2001
DNA Identification: Applications and Issues
As our understanding of human genetics has become more and more
sophisticated, scientists have developed increasingly powerful tools using genetic
information to aid in identifying individual people. DNA can be helpful in many
situations where identification is at question: For example, in criminal cases, forensic
DNA evidence can link a suspect, or a weapon such as a knife, to a crime scene. It
can also exclude a suspect. The best known sources of DNA evidence are blood and
semen, but increasingly it can be obtained from other items, such as a bottle cap, a
toothbrush, a bite mark on a piece of cake, or a fragment of a contact lens. Another
use is to identify human remains, such as of a soldier killed in battle. Or it might be
used to determine biological relationship, such as in paternity cases. This report
focuses on using human DNA to determine individual identity, but animal and plant
DNA can also be used, to identify species (for example, did a sample of meat come
from a protected whale species?) or even individuals (did a seed pod found in a truck
come from a particular tree at a crime scene?).1
Those and similar developments raise several policy issues, including the use of
genetic information in the criminal justice system, the impacts of continuing
technological improvements, and the effects of the technology on privacy and
individual rights. This report provides an overview of how genetic information is
used in identification and some of the issues associated with those uses. It begins by
discussing the unique properties of genetic information that make it a powerful tool
for identification and what is involved in making identifications from physical
evidence. Next is a description of current federal programs and activities related to
DNA identification, followed by discussion of issues raised by the development of thisth
new technology. Legislation activity in the 106 Congress focused on several
criminal-justice issues; for detailed discussion, see CRS Report RL30694, DNA
Evidence: Legislative Initiatives in the 106th Congress.
1 The examples in parentheses refer to real cases. The first involved testing whale meat
purchased in markets in Japan — C.S. Baker & S.R. Palumbi, “Which Whales Are Hunted?
A Molecular Genetic Approach to Monitoring Whaling,” Science 265 (1994): 1538. The
second was a 1993 murder case in which the DNA in seed pods in a defendant’s truck were
found to match the DNA of a palo verde tree at the crime scene — G. Sensabaugh and D.H.
Kaye, “Non-human DNA Evidence,” Jurimetrics Journal 38 (1998): 1-16.
Why DNA Can Be Used in Identification
The use of DNA in identification — sometimes called DNA typing or DNA2
profiling — results from its unique characteristics. Those characteristics also affect
how it can be used and the issues that arise from using it. Key features are described
It is a complex molecule, containing much information. DNA is a chemical,
deoxyribonucleic acid, consisting of modular components, called nucleotides, that are
connected in a linear sequence. Each nucleotide contains one of four bases — called
adenine, cytosine, guanine, and thymine, often designated by their initials A, C, G, and
T. Each DNA molecule consists of two complementary strands, with each adenine on
one strand paired with a thymine on the other, and each guanine with a cytosine. The
sequence of bases strung along a DNA molecule contains the information that forms
the basis of the genetic code of humans and most other organisms.3 Much of the
growing body of knowledge about DNA comes from efforts of the Human Genome
Project, a major goal of which is to create a map of the code and identify the sequence4
of bases on the DNA molecule. DNA identification technologies tap a small part of
the large set of information that the molecule contains.
Each person has billions of identical copies of DNA in the cells of the body.
Most of the billions of cells in a person’s body contain identical sets of DNA
molecules5 — approximately 3 billion base-pairs per cell altogether. Most of the
DNA is contained in 23 pairs of chromosomes in the nuclei of cells. Within each pair,
2 It is sometimes also called DNA testing. In this report, for clarity, that term is used only to
refer to medical tests.
3 For a more in-depth description, see Human Genome Program, U.S. Department of Energy,
Primer on Molecular Genetics (Washington, DC, 1992), [http://www.ornl.gov/hgmis/
4 The Human Genome Project is an international effort performed with both government and
private funding. The U.S federal effort began in 1988 with the signing of a memorandum of
understanding between the Department of Energy (DOE) and the National Institutes of Health
(NIH). At DOE, the program is housed in the Office of Biological and Environmental
Research (OBER) within the Office of Science (see the DOE genome project website at
[http://www.sc.doe.gov/production/ober/hug_top.html] for more information). At NIH, the
program resides in the National Human Genome Research Institute, NHGRI (see the NHGRI
website [http://www.nhgri.nih.gov:80/index.html] for more information). NHGRI was
established by statute, under the name National Center for Human Genome Research, in the
National Institutes of Health Revitalization Act of 1993 (P.L. 103-43). The DOE genome
efforts do not have separate statutory authorization.
5 The major exceptions are germ cells (eggs, sperm, and their precursors) which have 23
single, not paired, chromosomes, and red blood cells, which have no DNA. DNA extracted
from blood comes from white blood cells. The DNA sequences in different germ cells are not
identical, because each contains half the full DNA complement, partially mixed through a
process called recombination. Also, rare changes, called mutations, can occur in a sequence
through damage, errors in replication, or other means. Only mutations that occur in the germ
cell line can be inherited — they are called germ-line mutations; all others are called somatic
one chromosome is inherited from the mother and the other from the father. Some
DNA, inherited only from the mother, is also contained outside the cell nucleus, in
structures called mitochondria, which are small bodies with many copies in each cell.
Because the human body contains so many copies of DNA, even a very small amount
of body fluids or tissues, such as blood or skin, can yield useful identifying
The structure of the molecule, and therefore the information it contains, varies
from person to person. DNA sequences from any two people will be the same at
many points along the DNA molecule, but the overall nuclear DNA sequence of each
person is unique, except for identical twins, who have identical DNA.6 On a single
chromosome, individual alleles, which are discrete components of a DNA sequence,
are inherited intact. Each pair of chromosomes has many pairs of alleles7 (the total
number is as yet unknown), with one member of the pair inherited from the father and
the other from the mother. The region of a pair of chromosomes containing a pair of
alleles is called a locus. The DNA of the two alleles inherited at any particular locus
may have identical or different base-pair sequences.8 It is the substantial variation in
DNA among people that is characterized with the technologies used in DNA
The structure is inherited, so the DNA of close relatives is more similar than
that of distant relatives or unrelated people. Since half of a person’s DNA comes
from each parent,9 full siblings also share half their alleles, on average, and
grandparents and first cousins, one quarter. However, since a son receives his Y
chromosome from his father, all the alleles on that chromosome are identical to the
father’s and are inherited through the paternal line alone.10 Similarly, all alleles in
mitochondria are identical to the mother’s and are inherited through the maternal line11
alone. Therefore, the mitochondrial DNA of a male, for example, will be identical
to that of his maternal grandmother, and his Y-chromosome DNA identical to that of
his paternal grandfather. The inheritance patterns of DNA mean that analyzing a
person’s DNA can also provide identifying information about a relative.
Most of a person’s DNA has no known biological function. Genes are segments
of DNA that contain the code for making specific chemicals (mostly proteins).
Humans have tens of thousands of genes (the exact number is not yet known), but
6 except for mutations and certain other minor variations.
7 except for the sex chromosomes in males (see below).
8 The technical term is homozygous if they are identical and heterozygous if they are different.
9 Exceptions are mitochondrial DNA and the sex chromosomes in males, in whom the Y
chromosome is much smaller than the X and contains many fewer alleles.
10 The power of using this male-line inheritance of the Y-chromosome was demonstrated in
the genetic research that provided strong support for the assertion that Thomas Jefferson
fathered at least one son by Sally Hemings — Eugene A. Foster and others, “Jefferson
Fathered Slave’s Last Child,” Nature 396 (1998) 27–28.
11 except for mutations.
they comprise only a small part of a person’s DNA, most of which has no known
function. However, many nonfunctional segments of a person’s DNA can be
characterized with the techniques of molecular biology; those segments are called
markers. The distinction between genes and markers has implications for use in
identification that will be discussed later.
DNA is easily preserved with the structure intact. DNA is a surprisingly stable
chemical, and sequences can be easily preserved intact in dried or frozen samples.
However, DNA can be degraded, particularly in warm or moist environments or in
the presence of many common chemicals. The ability to use DNA in identification
depends both on the size and the condition of the sample. That has consequences for
sample collection, handling, and storage in applications such as law enforcement.
How DNA Is Used in Identification
Using DNA in identification requires comparing DNA whose source has not
been determined (such as from a crime scene or from a child in a paternity case) with
DNA whose source is known (such as from a suspect or from a putative father). Four
basic steps are involved: characterization, comparison, calculation, and interpretation.
They are each discussed below, along with implications for applications and issues.
The first step is to characterize or profile corresponding DNA sequences in the
samples to be compared. For forensic evidence in a criminal case, one or more
samples of blood, semen, or other sources of DNA related to the crime will be
processed, as will usually a sample from one or more suspects. It is not possible with
current technology to characterize the entire DNA sequence. Instead, alleles at
specific loci are characterized. The technique used depends on the particular kind of
marker. The major kinds of markers in use today are VNTRs, STRs, mtDNA, and
certain simpler sequences.
For most of the fifteen years during which DNA typing has been used in
forensics, VNTRs (variable number of tandem repeats) have been the major kind of
marker used. They have the greatest potential power to identify or exclude, because
they vary more from person to person than any other DNA system used in
identification. They consist of sections of DNA in which a sequence of about 15–70
bases is repeated many times. Analysis of a sample using VNTRs measures the
approximate number of repetitions in each marker examined. That number varies
substantially from person to person. Analysis of VNTRs uses RFLP12 technology,
which requires that substantially more DNA be present, and in good condition, in the
12 RFLP stands for restriction fragment length polymorphism. A chemical called a restriction
enzyme cuts the DNA molecule in certain (restricted) places corresponding to particular base
sequences. The result is pieces of DNA called restriction fragments that vary among people
— are polymorphic — in length. RFLP and VNTR are sometimes used interchangeably to
refer to the analysis of VNTRs using RFLP technology.
sample than do the other systems described below. Therefore, many samples that are
degraded or have very small amounts of DNA cannot be processed using VNTRs.
STRs (short tandem repeats) also consist of repeated sequences, but the number
of bases repeated is smaller (2–4), as is the total length of DNA comprising an allele
(approximately 100–300 bases for STRs versus 500–10,000 for VNTRs). STRs are
not as variable as VNTRs, and therefore more loci must be typed to obtain the same13
resolving power as with VNTRs. Nevertheless, STRs can be examined with much
smaller samples (for example, requiring a bloodstain the size of a pinhead rather than
one the size of a quarter coin), for two reasons. First, DNA-amplification procedures
(known as PCR — the polymerase chain reaction) can be used. The process uses the
ability of DNA to replicate itself to make many identical copies from a small initial
amount of DNA. The procedure does not yet work with VNTRs. Second, because
the sequences are shorter, they are less likely to be damaged if the sample is degraded.
Therefore, such samples can often yield more usable results with STRs than with
VNTRs. STRs also have some other technical advantages over VNTRs. In
particular, they can be processed in the laboratory much more quickly and they are
easier to interpret. For those reasons, STRs, which first became available for forensic
use a few years ago, are quickly replacing VNTRs as the standard marker for typing.
Mitochondrial DNA (mtDNA) can be used in even smaller or more degraded
samples than STRs. That is because the relevant DNA sequences are moderately
short (approximately 1,200 bases or less), there are thousands of copies per cell, and
they can be amplified. This kind of DNA can be extracted even from skeletal remains
and has been used, for example, in identifying the remains of armed forces personnel
from conflicts as far back as World War II.14 Because of the way it is inherited,
mtDNA from remains must be compared with samples obtained from maternal
relatives. Therefore, it cannot be used to distinguish among maternally related
persons. Also, mtDNA is not as variable as STRs and VNTRs, and it is therefore not
as powerful a tool for making positive identifications. It also takes much longer to
analyze than STRs.
Some other nuclear DNA markers, such as DQA and Polymarker, are also used
to aid in identification. Those consist of simpler DNA sequences than STRs or
VNTRs. They can be amplified and require only small amounts of DNA, but they are
less variable than STRs or VNTRs, and some of the loci used are closely linked to
genes or are genes themselves. However, they can be processed very rapidly and are
often used to determine quickly whether a potential source of DNA should be
eliminated or investigated further using more sensitive marker systems. They are also
used extensively in paternity cases.
13 Twelve STR loci provide about the same power to identify as five VNTR loci for most
populations (Dennis Reeder, National Institute of Standards and Technology, conversation
with the author, 9 June 2000). The current standard set of STR loci established by the
Federal Bureau of Investigation for use in law enforcement contains 13 loci.
14 Armed Forces DNA Identification Laboratory, “AFDIL…about us…mtDNA,”
[http://www.afip.org/oafme/dna/mtdna.htm], 22 October 1999.
Once DNA from the samples is characterized, the resulting profiles are
compared. If one or more alleles differ in the two samples, they are said not to match.
In that case, if the analysis was performed correctly, the samples did not have the
same origin. In the case of a criminal suspect, that means that the suspect did not
produce the DNA found in the evidence. In a paternity case, it means that the
putative father was not the biological father. If an attempt is being made to identify
remains using mtDNA, it means that the person whose remains were typed is not
related to the supposed maternal relatives. This use of DNA as a means of exclusion
is a powerful and important use, particularly in the criminal justice system. A
common estimate is that one-quarter of named suspects are excluded in cases where
DNA evidence is used.15 Also, in the past few years, more than 60 people convicted
of violent crimes in the United States have been subsequently exonerated as a result16
of DNA evidence.
If the profiles match — that is, if they are identical at every locus — then they
could have come from the same source. Alternatively, the source could be someone
else who has an identical profile for the markers that were examined. The way such
a match is treated in DNA evidence has usually been different than for other sources
of identifying information, such as fingerprints. In the latter, a match has usually been
considered either a positive identification or inconclusive — for example, either the
suspect left the fingerprints or could not be ruled out — depending on the quality of
the print. For DNA, in contrast, the science of population genetics provides in many
cases a way of estimating quantitatively the chances that the DNA could have come
from another source. The chances of such a coincidental match depend on both the
number of markers used and the variability exhibited by those markers. Therefore,
results of DNA analysis have usually been given in terms of the probability that such
a match could be coincidental rather than a conclusion about identity.17
For several years, there was controversy about how best to perform calculations
to estimate probabilities in DNA identification, but that issue is now largely, although
not completely, settled. The specific procedures can be complex and will vary
depending on the circumstances and the system used. However, there are two basic
15 Louis J. Freeh, Ensuring Public Safety and National Security Under the Rule of Law: A
Report to the American People on the Work of the FBI 1993 - 1998, Federal Bureau of
Investigation, 2000, 36, available at [http://www.fbi.gov/library/5-year/5YR_report_.PDF].
16 National Commission on the Future of DNA Evidence, Postconviction DNA Testing:
Recommendations for Handling Requests, National Institute of Justice, NCJ 177626
(September 1999), 2, available at [http://www.ojp.usdoj.gov/nij/pubs-sum/177626.htm].
17 However, those distinctions between interpretation of fingerprint and DNA analysis are
beginning to blur (see section on interpretation below).
The population frequency of each allele in the profile — that is, the percentage
of people who have that allele in the population examined — is identified. In most
cases, those frequencies are estimates drawn from extensive data banks such as those18
compiled by the FBI. Subpopulations within a country such as the United States
vary in the frequencies of different alleles, and therefore comparisons may be made
to data from the most relevant subpopulation, usually a racial or linguistic group, or,
if it is not known what group the source of DNA belonged to (as is often the case
with DNA from crime scenes), from two or more subpopulations.
The frequencies are multiplied together, with the aid of appropriate mathematical
formulas, to produce an estimated probability that someone drawn at random from
the population would have that profile. If enough loci are used, those probabilities
can be very small — on the order of one chance in billions or even trillions — even
though the frequency of any given allele is likely to be on the order of 1–10%.
In some cases, probability calculations may not be appropriate or useful. For
some loci, the frequencies of alleles in the population as a whole might not be known,
or there might be too few loci or too few alleles at a locus to yield useful probability
estimates. In such a case, a match means that a putative source cannot be excluded,
but there might be many other potential sources. Or the particular case might not
require a calculation. For example, if the question is which of two men fathered a
child, no calculation is necessary if the profile of one yields a match and the other does
not. Or, in a case of identification of remains with mtDNA, it might be known that
a soldier came from one of four families. If the maternal mtDNA profile of only one
of the four yields a match, then the soldier came from that family. However, if more
than one source produces a match in either of those two examples, then DNA
evidence cannot resolve the question unless more loci can be examined. Another
situation that can present problems for making useful calculations is where DNA from
more than one person is in the sample. In such mixed samples, if the DNA
contributed by different people cannot be separated,19 then it might only be possible
to determine if someone can be excluded.
There are two essential questions involved in interpreting the results of a DNA
identification: Did the DNA come from the person (or family) who is thought to be
its source, and what is the significance of the answer for the case at hand?
The probability calculation, when performed, is used to help answer the first
question. If the probability estimate is low enough, the expert who provides it may
18 See, for example, Federal Bureau of Investigation, VNTR Population Data: A Worldwide
Survey (5 volumes), (Quantico, VA: FBI Academy, 1993). The data are drawn from
anonymized samples from blood banks or other sources. They should not be confused with
the databanks housing DNA profiles of convicted criminals, which are discussed below.
19 For example, if the DNA comes from different kinds of tissues or fluids, as in a vaginal
swab from a sexual assault, it is often possible to separate the male and female DNA.
However, if DNA from more than one male is present, separating those is often not possible.
declare a positive identification — that the person in question was in fact the source
of the DNA.20 However, the question of what probability level (or other criterion) is
required to ensure a positive DNA identification is not yet settled, and in many cases,
experts may be reluctant to make such a declaration, preferring to provide simply the
probability estimate for the jury or other trier of fact to interpret. That is in contrast
to the situation with fingerprints, which are accepted as unique (even in identical
twins, unlike DNA) but for which probabilities are not calculated. Fingerprints have
been used effectively for much longer than DNA, so little question remains about how
they can provide positive identification. In addition, they are not as amenable to
statistical analysis as is DNA evidence, although that is beginning to change as
automated fingerprint analysis systems are developed and refined. However, with
DNA, triers of fact may interpret a very low probability estimate as a positive
identification, provided that there are no significant questions about how the evidence
was handled, the quality of the laboratory analysis, or special circumstances of the
Once a positive identification has been made (or found very likely), its
significance must be determined. That would seem straightforward but is not always.
For example, in a rape case involving DNA evidence, if the probability of a
coincidental match is very small, that would be strong evidence of a sexual encounter
but would not of itself be proof of guilt, since the encounter might have been
consensual. However unlikely, it is also possible, for example, that there was a
laboratory error or even that blood or other sources of DNA might have been planted
by someone wishing to frame a suspect. For those and other reasons, a match
probability in a criminal case should not be interpreted as a probability of guilt.21
20 For example, in 1997, FBI experts switched from simply providing probability estimates
to stating in addition that a person whose profile matches a sample is the source of the sample,
provided that the calculated match probability is less than about 1 in 260 billion (Freeh,
Ensuring Public Safety, 35; Jennifer Smith, Laboratory Director, DNA Unit 1, Federal
Bureau of Investigation, “Comments regarding R&D Report,” Proceedings, National
Commission on the Future of DNA Evidence, 9 April 2000, [http://www.ojp.
usdoj.gov/nij/dnamtgtrans9/trans-e.html]). For comparison, that probability is about 1/50 the
chance that a person drawn at random from the entire world population would have that
profile. The original scientific work that provided the basis for the use of fingerprints in
identification established that the probability of a second person having the same fingerprint
pattern on a given digit was about 1/40 the reciprocal of the world population at the time —
National Research Council, The Evaluation of Forensic DNA Evidence, (Washington, DC:
National Academy Press, 1996), 57. The average match probability yielded by using 12 of
the 13 core STR markers used by the FBI is about 1 in 700 billion (James F. Crow, “Research
and Development Working Group Report,” Proceedings, National Commission on the Future
of DNA Evidence, 28 February 1999, [http://www.ojp.usdoj.gov/nij/ dnamtgtrans4/trans-
c.html]). Nevertheless, many forensic experts prefer to provide probabilities but not state firm
conclusions about the source of the DNA.
21 Such misinterpretations are sometimes called “the prosecutor’s fallacy.” This can be a
surprisingly complex issue, and a technical discussion of it is beyond the scope of this report.
For more information, see NRC, Evaluation of DNA Evidence, 133, 198; and David H. Kaye
and George F. Sensabaugh, Jr., “Reference Guide on DNA Evidence,” in Reference Manualnd
on Scientific Evidence, 2 ed. (Washington, DC: Federal Judicial Center, 2000), 539,
However, it does provide evidence for the jury or judge to consider in determining
whether the accused was the perpetrator of the crime.
Databases or indexes used in DNA identification are of three basic kinds. One
provides the allele frequencies that are used in calculations to estimate profile
frequencies and match probabilities. Such population databases are drawn from
anonymous samples and are separated or stratified according to the population group
(usually based on ethnicity or race) of the donors, since allele frequencies, and
therefore match probabilities for different profiles, may differ among such groups.
Sources are various, such as blood donors or medical patients.22
The second kind contains profiles of persons whose identity is known. One
example is databases with STR or VNTR profiles of convicted felons or of victims of
unsolved crimes (such as the CODIS database system, discussed below). Another is
databases with profiles of missing persons or their biological relatives, such as
mtDNA profiles of maternal relatives of armed forces personnel lost in past military
conflicts (see below). When a profile is obtained from a relevant sample whose
source is not known, the database can be searched to determine if a match, called a
cold hit, is found. For example, in an increasing number of cases, a suspect is
identified when a DNA profile from a crime-scene sample is searched against a
database containing profiles of persons convicted of violent crimes or other felonies.
The third kind of database contains profiles of persons whose identity is not
known. Samples might come from crime scenes or unidentified remains. When a
profile is obtained from a relevant sample whose source is known, the database can
be searched for cold hits, as above. For example, a profile obtained from a suspect in
another crime can be searched against a forensic database of profiles associated with
crimes for which there are no suspects. In 1999, a DNA profile of a male murdered
in Florida was found to match DNA evidence from nine rapes, three in Florida and six
in Washington, D.C.23
available at [http://air.fjc.gov/public/fjcweb.nsf/pages/16].
22 Such so-called “convenience samples” might seem unlikely to generate accurate frequencies
of the distribution of alleles in the underlying population, but they work surprisingly well
(NRC, Evaluation of DNA Evidence, 126).
23 Federal Bureau of Investigation, “First ‘Cold’ Hit Recorded in National DNA Index
System!”, Press Release, 21 July 1999, [http://www.fbi.gov/pressrm/pressrel/pressrel99/
coldhit.htm]; Kathleen Sweeney, “DNA testing links rapes to slain man,” Florida Times
Union, Thursday, 22 July 1999, Sec. A, 1. The index system is discussed below.
Federal Agency Programs and Activities
Federal agencies with significant involvement in DNA identification activities
include the Federal Bureau of Investigation (FBI), the National Institute of Justice
(NIJ), and the Bureau of Justice Assistance (BJA) in the Department of Justice; the
National Institute of Standards and Technology (NIST) in the Department of
Commerce; and the Armed Forces Institute of Pathology (AFIP), the Army Central
Identification Laboratory, Hawaii (CILHI), and the Army Criminal Investigation
Laboratory (USACIL)in the Department of Defense. Major activities are described
Federal Bureau of Investigation
Most forensic DNA evidence is developed and used by local or state law
enforcement agencies. However, the FBI provides many important services to those
agencies and is responsible for processing DNA evidence for cases under federal
FBI Laboratory. Major activities of the FBI Laboratory include training of
federal, state, local, and foreign law enforcement and crime laboratory personnel;
research and development in DNA typing technologies; development of an integrated
national DNA database program; and providing expert testimony in the courts.24 The
laboratory was the first public crime laboratory in the United States to perform
analysis of forensic DNA evidence, creating its DNA Analysis Unit in 1988. It is
currently the only such laboratory performing mtDNA analyses.
The laboratory also administers the Combined DNA Index System. CODIS is
a distributed system of local, state, and national DNA databases that are linked
electronically, permitting the comparison of profiles stored in different locations.
Begun as a pilot program in 1990, it was authorized in the DNA Identification Act of
1994 (P.L. 103-322). More than 40 states now participate. CODIS has several
indexes. One, a convicted offenders index, contains DNA profiles of persons
convicted of qualifying crimes. Current law does not specify qualifying federal
crimes. However, section 811(b)(2) of the Antiterrorism and Effective Death Penalty
Act of 1996 (P.L. 104-132) required that states, to be eligible for grants to improve
their capacity to perform forensic DNA analyses and certain other activities,25 collect
DNA samples from persons convicted of sexual felonies. All 50 states now require
24 See the laboratory Web site at [http://www.fbi.gov/programs/lab/labhome.htm] for more
25 Specifically, the grants are made “to carry out all or part of a program to establish, develop,
update, or upgrade…computerized identification systems…, the capability to
analyze…DNA…, and automated fingerprint identification systems…,” provided that they
are compatible with the relevant corresponding FBI systems (Sec. 811(b)(1)). Those State
Identification Systems formula grants are administered through the Bureau of Justice
samples from such persons. Most also collect samples from persons convicted of
murder or other violent crimes, and several from those convicted of any felony.26
Other indexes are a forensic index, which contains profiles of DNA samples
taken from crime scenes (especially from cases without any suspects); a population
file, which contains information on allele frequencies to be used in calculating match
probabilities; and a missing persons index containing profiles from unidentified
The national component of CODIS, called NDIS, the National DNA Indexing
System, has been in operation since 1998. Laboratories in 24 states currently27
contribute DNA profiles to NDIS. While CODIS initially used VNTR markers, they
are largely being replaced with the more powerful STRs (see section on sample
backlogs below). NDIS uses thirteen core STR markers that have been established
as a standard set by the FBI.
DNA Advisory Board. DNA typing is technology intensive. That makes issues
of quality control and assurance especially important. To help address such issues,
P.L. 103-322 required that the FBI Director establish a DNA Advisory Board (DAB)
to recommend quality assurance standards for forensic DNA analysis. Following
submission of DAB recommendations, the director established standards effective
October 1, 1998. They replaced standards that had been established by the Technical
Working Group on DNA Analysis Methods (TWGDAM), a practitioners’ group
representing federal, state, and local forensic laboratories and supported by the FBI.
As of 1998, most of the publicly funded crime laboratories in the United States
followed DAB or TWGDAM standards, and about half were accredited by an official28
organization. DAB is scheduled to dissolve at the end of 2000, at which time its
functions will be transferred to the renamed TWGDAM, now called the Scientific
Working Group on DNA Analysis Methods (SWGDAM).
National Institute of Justice
The National Institute of Justice (NIJ), a research agency within the Office of
Justice Programs, engages in several kinds of activities related to DNA evidence. The
institute supports research to improve speed, reliability, and sensitivity of DNA
profiling and to reduce its cost. There are three main activities, administered through
the NIJ Office of Science and Technology: the DNA Five Year Research Program,
the Forensic DNA Laboratory Improvement Program, and the National Commission
26 Dwight E. Adams, Statement, Legislative Hearing on H.R. 2810, the “Violent Offender
DNA Identification Act of 1999", H.R. 3087, the “DNA Backlog Elimination Act”, and H.R.
House Committee on the Judiciary, 13 March 2000, [http://www.house.gov/judiciary/
27 Adams, Hearing Statement.
28 Greg W. Steadman, Survey of DNA Crime Laboratories, 1998, Bureau of Justice Statistics
Special Report NCJ 179104, February 2000, 3, available at [http://www.ojp.usdoj.gov/bjs/
on the Future of DNA Evidence. The DNA Five Year Research Program
(1999–2003), is awarding $5 million per year to support research aimed at reducing
the cost and time of processing DNA samples, developing technologies to enhance
the reliability of DNA analysis and perform analyses at crime scenes, and developing
standard test materials and new markers.29
The Forensic DNA Laboratory Improvement Program was initiated in FY1996
with DNA Identification Grants, authorized in P.L. 103-322 (Sec. 210302).
Appropriations for the program have grown annually (see table). From FY1996–
FY1999, the program awarded grants to state and local governments, up to 75% of
the total cost of the project, to develop and improve the abilities of forensic
laboratories to analyze DNA evidence. In FY2000 and FY2001, funds were awarded
under a new series of grants authorized by the Crime Identification Technology Act
of 1998 (P.L. 105-121), with half allocated by NIJ to the DNA Identification Program
and half to the elimination of sample backlogs (see below). That program can fund30
up to 90% of the total cost of a project.
Appropriations for State and Local DNA Laboratory Support,
(in millions of current dollars)
Note: FY2000 includes funds for addressing sample backlogs. Source: FY1996, P.L. 104-134;
FY1997, 104-208; FY1998, P.L. 105-119; FY1999, P.L. 105-277; FY2000, P.L. 106-113; FY2001,
A 1996 NIJ-funded report31 led Attorney General Reno to charter, in September
29 National Institute of Justice, Technology Development Portfolio: Investigative and
Forensic Sciences, [http://www.ojp.usdoj.gov/nij/ sciencetech/invest.htm], 16 January 2000.
30 Information on NIJ grant programs can be found on the agency’s Web site at [http://www.
31 Edward Connors and others, Convicted by Juries, Exonerated by Science: Case Studies
in the Use of DNA Evidence to Establish Innocence After Trial, National Institute of Justice
Research Report, NCJ 161258 (June 1996), [http://www.ncjrs.org/pdffiles/dnaevid.pdf].
the DNA Commission). This four-year NIJ commission is examining several topics,
including the postconviction use of DNA, legal concerns, training and technical
assistance, and future technological developments. Its members include
representatives from federal, state, and local law enforcement agencies, the judiciary,
defense lawyers, and other groups and areas of expertise. Rather than producing a
single final report, the commission is developing reports on specific issues and making
recommendations on an ongoing basis. In 1998, the commission first identified the
need for a special effort to address a backlog of hundreds of thousands of DNA
samples, taken from convicted persons and from crime scenes, that have not yet been
processed (see section on sample backlogs below).32 A 1999 report33 recommended
procedures for handling postconviction DNA-typing requests.
Bureau of Justice Assistance
The Bureau of Justice Assistance supports state and local criminal justice
programs. It administers formula grants, including Byrne grants (42 U.S.C. 3751),
which can be used, among other purposes, to develop or improve the DNA-analysis
capabilities of forensic laboratories; and State Identification Systems grants, which can
be used, among other purposes, to help laboratories develop their capabilities with
respect to CODIS.34
National Institute of Standards and Technology
The National Institute of Standards and Technology (NIST) has played a major
role in the development of standards for DNA profiling, as part of its long-standing35
role in developing technological standards and measurements generally. The
standard reference materials that NIST has produced, both for VNTR and STR36
analysis, are used by laboratories to test the accuracy of their analyses. They are
therefore important components of quality-assurance and control activities. NIST
also performs research on new DNA-typing technologies, both through the
Biotechnology Division (within Scientific and Technical Research and Services) and
the Advanced Technology Program (within Industrial Technology Services). Many
of the DNA-forensic activities in which NIST engages are performed in collaboration
32 Paul Ferrara, “CODIS Backlog Elimination Report,” Proceedings, National Commission
on the Future of DNA Evidence, 8 June 1998, [http://www.ojp.usdoj.gov/nij/dnamtgtrans2/
trans-f.html]. Specific actions to eliminate the backlog were later recommended to the
33 DNA Commission, Postconviction DNA Testing.
34 Information on BJA grant programs can be found on the agency’s Web site at
[http://www.ojp.usdoj.gov/BJA/html/fund1.html]. For information on Byrne grants, see also
Garrine P. Laney, Crime Control Assistance Through the Byrne Program, CRS Report 97-
35 For general information on NIST, see Wendy H. Schacht, The National Institute of
Standards and Technology: An Overview, CRS Report 95-30, 6 July 2000.
36 For more information on STRs, see John M. Butler and Dennis J. Reeder, “Short Tandem
Repeat DNA Internet Database,” [http://www.cstl.nist.gov/biotech/strbase/], 24 May 2000.
with other agencies, such as the NIJ and Department of Defense, and with private
Department of Defense
The two major uses of DNA identification by the Department of Defense (DOD)
are in criminal investigation (see below) and in the identification of remains of military
personnel. To aid in such identification, DOD established in 1988 the Office of the
Armed Forces Medical Examiner within the Armed Forces Institute of Pathology,37
under the Assistant Secretary of Defense for Health Affairs. DNA was first used to
identify combat fatalities in 1991 in Operation Desert Storm. At that time, the DNA
Registry38 was formed, consisting of two components. The Armed Forces Repository
of Specimen Samples for the Identification of Remains (AFRSSIR) collects and stores
DNA samples taken from active duty and reserve personnel.39 Armed Forces
personnel are required to provide DNA samples for deposit in the repository.
Currently, the repository holds more than 3 million specimens. When a casualty
occurs that requires DNA identification, the Armed Forces DNA Identification
Laboratory (AFDIL) processes relevant samples to produce nuclear (STR and other
PCR-based systems) or mtDNA profiles. AFDIL also performs analyses of samples
in other selected cases.40 Two examples of the latter are assisting the FBI in
identifying remains after the fire at the Branch Davidian compound in Waco, Texas,
in 1993, and assisting the National Transportation Safety Board in identifying
passengers who died in the crash of TWA Flight 800 in 1996.
Another DOD unit involved in DNA identification of remains is the United States
Army Central Identification Laboratory, Hawaii (CILHI).41 The laboratory recovers
and identifies remains of military personnel lost in past conflicts and unaccounted for.
AFDIL performs mtDNA analysis on the recovered remains. AFDIL maintains a
database of mtDNA profiles from maternal relatives who volunteer to provide
samples, for matching against profiles from remains. As of June 1999, 154 matches
had been obtained (including 112 from Vietnam and 34 from World War II). The
most prominent was the identification of First Lieutenant Michael J. Blassie, USAF,
whose remains had been interred in 1984 as the Vietnam Unknown in the Tomb of the
Unknowns in Arlington National Cemetery. Lt. Blassie had been lost in 1972 in
Vietnam. Samples of mtDNA from maternal relatives in seven families involved in the
37 See Department of Defense, “Armed Forces Institute of Pathology (AFIP),” DOD Directive
38 See Department of Defense DNA Registry, “Welcome to…,” [http://www.afip.org/oafme/
dna/history.htm], 22 October 1999.
39 The target date for completion of collection from reserve personnel is December 2002
(AFRSSIR, “Repository History,” [http://www.afip.org/oafme/dna/History.htm], accessed
27 June 2000). Specimens may be retained for up to 50 years. However, personnel may
request destruction of specimens once they have completed their service.
40 For guidelines, see “AFDIL...DNA Services,” [http://www.afip.org/oafme/dna/
outsidedna.htm], 22 October 1999.
41 See “The United States Army Central Identification Laboratory, Hawaii,” [http://www.
cilhi.army.mil/], 19 June 2000.
investigation were compared with mtDNA from a bone fragment taken from the
tomb, and the samples from the Blassie family provided a very close match.
The United States Army Criminal Investigation Laboratory (USACIL) performs
DNA analysis for criminal investigative agencies within the Department of Defense.
The laboratory analyzes approximately 500 cases per year and recently converted
from VNTR to STR analysis using the 13 core loci. USACIL participates in CODIS,
providing profiles for NDIS of evidentiary samples for specific cases and searching
for matches in cases under investigation, where appropriate. However, the armed
forces do not currently collect samples from convicted offenders (see the section
below on issues). USACIL works only on cases with a military connection.42
Three other federal agencies support research that has contributed to the
scientific basis for advances in DNA identification. The Office of Biological and
Environmental Research, in the Department of Energy, and the National Institutes of
Health, in the Department of Health and Human Services, are the lead agencies for
the Human Genome Project. They also support research on ethical, legal, and social
aspects of DNA identification and other applications of advances in genomics.43 The
National Science Foundation, an independent agency, supports relevant basic research
in molecular biology and in the social sciences at universities and other research44
Relevant Public Laws
When DNA evidence first became available in the United States, questions often
arose about the quality of analyses and the absence of widely accepted standards.45
Also, the typing of specimens was expensive,46 and its availability to many state and
local law enforcement agencies was therefore uneven. The DNA Identification Act
of 1994, a subtitle of the Violent Crime Control and Law Enforcement Act of 1994
(P.L. 103-322), addressed quality control and privacy issues. It authorized the DNA
42 This description is based on information provided by Larry Chelko, Director, USACIL,
email communication with the author, 25 August 2000.
43 Details can be found at the program Web sites — for DOE, see [http://www.ornl.gov/
hgmis/elsi/elsi.html], and for NIH, see [http://www.nhgri.nih.gov/About_NHGRI/Der/Elsi].
44 For general descriptions of research programs in these agencies, see the following CRS
reports: Richard E. Rowberg, Department of Energy Research and Development Budget for
FY2001: Description and Analysis, CRS. Report RL30445, 12 September 2000; Pamela
Wolfe Smith, The National Institutes of Health: An Overview, CRS Report 95-96, 15
September 2000; Christine M. Matthews, U.S. National Science Foundation: An Overview,
CRS Report 95-307, 20 September 2000.
45 National Research Council, DNA Technology in Forensic Science, (Washington, DC:
National Academy Press, 1992), 97–110.
46 Ibid., 153–154.
Identification Grants program administered by NIJ (see above). The law authorized
appropriations for the program through FY2000. It also authorized use of Drug
Control and System Improvement Grants for similar purposes (42 U.S.C. 3751; those
grants are part of the Edward Byrne Memorial State and Local Law Enforcement
Assistance Programs, or Byrne grants). The DNA Identification Act required that
forensic laboratories receiving the grants follow specified quality assurance and
privacy provisions. Recipients are to meet current quality-assurance standards, as
specified by the FBI director, and undergo regular proficiency testing in DNA
analysis. The use of DNA samples analyzed by recipients are restricted to use in law
enforcement, judicial proceedings, and criminal defense. However, anonymized
samples can also be used in population databases, research, and quality control
The act also provided for the establishment of the FBI’s DNA Advisory Board,
to recommend quality-assurance and proficiency-testing standards to the director
(Sec. 210303), and required that FBI personnel engaging in DNA analysis undergo
regular proficiency testing (Sec. 210305). It also authorized the establishment of
CODIS indexes containing profiles of persons convicted of crimes and samples
recovered from crime scenes or unidentified remains (Sec. 210304). The Consolidated
Appropriations Act of 2000 (P.L. 106-113) additionally provided for an index of
profiles from “samples voluntarily contributed from relatives of missing persons” (Sec.
The Antiterrorism and Effective Death Penalty Act of 1996 (P.L. 104-132)
authorized the application of CODIS to federal crimes and those committed in the
District of Columbia (Sec. 811(a)(2)). It also authorized grants to state and local
government for participation in CODIS. To be eligible, states must collect DNA
samples from persons convicted of felony sex crimes (see above). Also, the National
Institutes of Health Revitalization Act of 1993 (P.L. 103-43) established as one of the
purposes of the National Human Genome Research Institute “reviewing and funding
proposals to address the ethical and legal issues associated with the genome project.”
The Crime Identification Technology Act of 1998 (P.L. 105-521, 112 Stat.
Information, and Communication. Grants can be awarded for a broad range of
activities to, among other things, improve state capabilities in crime identification and
promote compatibility and integration among local, state, and federal identification
systems, and including accreditation and certification programs relating to DNA
analysis. Funding is authorized through FY2003.
Current and Emerging Issues
Some of the issues discussed below, particularly those related to law
enforcement, were the subject of legislation in the 106th Congress. See CRS Report
RL30694 for a discussion of those legislative initiatives.
Law Enforcement and Criminal Justice
Sample Backlogs. DNA samples are now collected in all 50 states from persons
convicted of certain crimes. Also, DNA evidence is routinely gathered from crime
scenes and victims in cases of rape, murder, and other violent crimes. As DNA
technology has improved and states have increased its scope in criminal justice
activities (see next section), the workload in forensic laboratories has increased. For
example, from 1996 to 1997, casework increased 40%, from a total of about 15,000
to 21,000, and convicted-offender samples increased more than 60%, from 72,000 to47
116,000. Given limited resources, forensic laboratories must prioritize the analysis
of those and other samples they receive. Therefore, many samples, such as from
scenes of crimes for which there are no suspects, and from convicts who are not
suspects for additional crimes, are given lower priority. Currently, there is a backlog
in the United States of several hundred thousand such samples that have not been
analyzed and entered into CODIS databases. That means that profiles from those
samples are not available for database searches. Given that “cold hits” have been
identified through such searches, the DNA Commission and others have proposed that
funding be increased to process that backlog. Also, the $30 million appropriated by
Congress in FY2000 for DNA grants included funding for processing backlogged
An additional factor is the conversion of existing profiles from VNTRs to STRs.
Some states have not converted to using the core STR markers, and profiles that use
other markers cannot be searched against NDIS or CODIS STR records. Conversion
requires that a DNA sample be retyped — reanalyzed with STR technology. More
than 200,000 existing profiles may need such retyping. Also, there are many “owed”
samples — that is, samples that can be taken under existing law but have not been.
Those include many paroled or released convicts. The number of such owed samples
nationwide may exceed the number currently backlogged.48
Failure to process backlogs may have several consequences. Crimes that might
be solved with the help of a database match may remain unsolved. That is of
particular concern in cases where a perpetrator is likely to perform additional crimes,
or where a database match would prevent an innocent person from being wrongly
suspected or perhaps even charged with the crime. Also, crime-scene samples from
unsolved crimes may eventually be destroyed as statutes of limitations expire,
permanently eliminating any possibility of typing any DNA evidence.
The size of the backlog makes the question of prioritization particularly
important, but it is not straightforward. For example, typing a person’s DNA is
usually much less expensive than typing DNA from a crime scene, because in the
latter case, several samples must usually be processed and, unlike with convicted-
offender samples, processing cannot usually be automated with existing technology.
That might suggest that the highest priority should be given to typing convicted
47 Steadman, Survey, 6.
48 See “CODIS Offender Database Backlog Reduction Discussion,” Proceedings, National
Commission on the Future of DNA Evidence, 23 November 1998, [http://www.ojp.usdoj.
persons, since many more samples could be processed per dollar invested. However,
to make a successful cold hit requires that crime-scene samples also be typed, so that
they can be compared with the profiles of convicted persons. Also, logistically, it is
often easier to obtain a sample from someone newly in custody than from someone
who has been released. However, a person while in prison is not a threat to the
community, whereas someone who has been released might be; nevertheless, a
prisoner might be discovered through DNA analysis to have committed other crimes
before having been imprisoned. Finally, solving old crimes with the help of DNA
evidence serves justice and might help prevent future crimes, but it might also have
negative consequences. For example, reopening a case that is several years old might
retraumatize victims or their families who have worked to recover from the effects of
Additional federal funding to help reduce the backlog could substantially increase
the speed with which backlogged samples are processed, easing uncertainties about
how best to prioritize the samples and increasing the rate at which crimes are solved.
Also, the existence of the backlog could be attributed in part to the success of the
DNA Identification Act grant program, and the grant-eligibility requirements in the
Antiterrorism and Effective Death Penalty Act of 1996 discussed above. In addition,
the processing of backlogs would have benefits across states, especially with respect
to “travelling offenders” who commit crimes in more than one state.
Databases. A central issue relating to the kinds of crimes for which DNA is
collected from convicted persons for inclusion in a profile database or index. States
vary in the crimes for which they collect DNA samples. Qualifying offenses include,
at a minimum, felony sex crimes, but several others are included by different states:
offenses against children (40 states), murder (36), assault and battery (27), kidnapping
(22), robbery (19), burglary (14), and all felonies (6). Also, 24 states collect samples
from juveniles convicted of qualifying offenses, most collect retroactively from
incarcerated convicts, and some collect from those previously paroled or on49
probation. In the United Kingdom, the Forensic Science Service, an executive
agency of the Home Office, maintains a national database with DNA profiles of
“suspects charged, reported, cautioned or convicted for a recordable offence,” which
is any crime punishable by imprisonment, plus certain other specified offenses.50
Profiles are removed if a suspect is exonerated or acquitted. As of July 2000, the
U.K. database held approximately 700,000 profiles and had yielded more than 77,000
matches of suspects to crime scenes since its inception in 1995, with almost 130,000
profiles removed following acquittal.51
49 Adams, Hearing Statement. Figures cited are as of December 1999 from an FBI survey of
laboratories participating in CODIS. States have continued to pass laws expanding the list
of qualifying crimes, so the numbers cited for some categories have since increased.
50 Among the additional recordable offenses are public drunkenness, taking part in a
prohibited assembly, and certain kinds of illegal hunting — Government of Great Britain, The
National Police Records (Recordable Offences) Regulations 2000, Statutory Instrument
51 The Forensic Science Service, Annual Report and Accounts 1999–2000, 25 July 2000,
In determining whether to broaden the range of qualifying crimes, questions
might be raised such as the following: What is the cost-effectiveness of profiling
those convicted of a particular class of crime? For example, are the resources needed
to profile those convicted of nonviolent crimes more effective if spent on profiles or
on other aspects of crime solving and prevention?52 What is the proper balance
between using profiles to help protect citizens from crime, on the one hand, and the
need to protect the civil liberties and privacy of those who might be subject to
profiling, on the other? The power of DNA evidence has led some to call for indexing
profiles of all arrestees, or even of all citizens at birth, while others have raised
concerns about the effects of such measures on civil liberties.53
Postconviction Analysis. A powerful use of DNA evidence is in exonerating
the innocent, including in some cases where people have been wrongfully convicted.
Several issues are associated with postconviction DNA analysis. Among them are
when such procedures are appropriate, and what is the proper role of the federal
There are several potential reasons why DNA analysis might not have been done
when a case was first prosecuted. One is that the identity of the suspect might not
have been an issue at the trial — for example, the involvement of a suspect in a sexual
encounter might not have been in question, but rather whether the encounter was
consensual. In such a case, DNA evidence would probably not be relevant. Another
possibility is that no DNA evidence was found at the time, but turned up later. A
third is that there was DNA evidence, but the technology to analyze it was not
available when the case was prosecuted. The first use of DNA typing in the United
States was in the late 1980s, and it was not widely available until a few years ago.
available at [http://www.forensic.gov.uk/forensic/corporate/annual_rep/
52 The answer to these two questions depends on several factors. One is the frequency with
which those who commit nonviolent crimes also commit other crimes that DNA evidence
would help to solve. There are few data on this topic so far, but the experience of Virginia
might be illustrative. Using a convicted-offender database of approximately 118,000 profiles
as of June 2000, Virginia obtained 156 matches of offenders to crimes. The state collects
DNA samples from everyone convicted of a felony. Dr Paul Ferrara, Director of the Virginia
Division of Forensic Sciences, has estimated that about half of those “hits” would not have
occurred had the database been limited to violent offenders (personal communication with the
author, 28 August 2000). Another factor is the cost of collecting samples and analyzing the
DNA for a convicted offender. The DNA Commission has estimated the latter at
approximately $50 per sample with current technology, but it can vary significantly depending
on circumstances. A third factor is cost-effectiveness of other methods of crime solving and
prevention. These and other factors involved can be difficult to measure and compare
53 Rose Marie Arce, “Surveillance and DNA testing are among the latest police weapons. But
how will we balance fighting crime and preserving civil rights?” Newsday, 30 May 1999, sec.
A, 17. See the section on privacy below for further discussion of this issue.
In some cases, relevant DNA evidence might have been found and analyzed at
the time of the original prosecution but yielded inconclusive results because of the
limitations of the technology at the time. The much more sensitive STR technology
did not become an established standard until the late 1990s, and many states still use
the older, less sensitive VNTR markers. STRs or the potentially even more sensitive
(but less powerful) mtDNA can provide useful profiles from much smaller or more
degraded samples than VNTRs. Consequently, an analysis that was inconclusive with
VNTR technology could lead to either a definitive exclusion or, alternatively, a strong
match if a more sensitive system is used. Improvements in DNA technology are likely
to continue into the future, potentially making even more samples amenable to typing.
Most states currently do not permit new trials, based on newly discovered
evidence, more than three years after conviction.54 However, DNA evidence,
properly handled, is very stable and can often provide useful information even ten
years or more after it was initially deposited.55 This confluence of advances in the
technology, the stability of DNA evidence, and its strong and growing exclusionary
power raise questions such as whether the time limit should be extended for cases
where DNA evidence newly discovered or analyzed can be probative, and whether
postconviction legal procedures should be changed to accommodate the particular56
features presented by DNA evidence.
To date, a few states have passed laws that specifically permit postconviction
DNA analysis for convicted persons claiming actual innocence. In considering
proposed federal legislation, Congress may consider questions including the
following: Should such legislation apply only to crimes under federal jurisdiction, or
should it also apply to states? What evidence standards should a petitioner be
required to meet for the procedure to be permitted? Should there be a requirement
for evidence to be preserved beyond exhaustion of appeals, and if so, for how long
should it be preserved? Under what circumstances should elimination samples from
victims or other third parties be required? What, if any, accommodation should be
made for retyping when new advances are made in DNA technology? Should the
government pay for analysis requested by an indigent inmate? Should wrongfully
convicted persons receive compensation?57
54 DNA Commission, Postconviction DNA Testing, 9.
55 In some cases, convicted persons who were actually innocent have served more than ten
years before being exonerated by DNA evidence. For examples of those and other cases of
postconviction exoneration, see Edward Connors and others, Convicted by Juries, Exonerated
by Science: Case Studies in the Use of DNA Evidence to Establish Innocence After Trial,
National Institute of Justice Report NCJ161258, June 1996.
56 DNA Commission, Postconviction DNA Testing, 10. For example, evidence containing
DNA would usually not be newly discovered but would likely already be in the control of the
prosecution; however, a DNA analysis might not have been done or might have been
inconclusive because of earlier technological limitations. Also, some law enforcement
agencies destroy evidence in a case after all appeals are exhausted (Connors, Convicted by
57 See Eric A. Fischer, DNA Evidence: Legislative Initiatives in the 106th Congress, CRS
Report RL30694, 12 January 2001, for discussion of these issues.
As more powerful DNA-typing techniques have become more affordable, their
use in paternity analysis has grown. An emerging issue involves cases where men
who have legally acknowledged paternity, sometimes for several years, are shown by
DNA typing not to be the father of the child. A central question is whether the men
in such cases, who may claim not to be responsible for child support since they are not
the child’s biological father, should nevertheless continue to be held responsible. A
concern of child advocates is the potentially negative impacts on children if such
challenges are allowed. States have varied in their treatment of this issue.58
Future Directions of the Technology
Current research on improving DNA identification is performed or supported
mainly by NIST, NIJ, the FBI Laboratory, AFDIL, and private industry. Such
research has several broad goals. Improvements in sensitivity of analyses would make
useful typing possible from even smaller or more highly degraded samples than at
present. That could permit more powerful identification (or exclusion) from DNA in
trace evidence from, for example, saliva, tears, and skin cells. Coupled with further
improvements in specificity, such as via use of a larger set of markers, it would also
increase the ability of DNA evidence to make positive or unique identifications. Even
identical twins possess minor genetic differences that could eventually be detected.
Reducing the time required to complete typing could lead to more timely following
of leads as well as more quickly removing from further consideration those persons
excluded by the evidence. As the costs of performing DNA analysis has come down,
its use has increased. Further reducing cost would make use of DNA identification
more widely available. Research on automation and miniaturization of the typing
process may lead to cost reductions and further improvements in quality control.
Conceivably, an automated, portable DNA-typing system could eventually be
developed that would permit on-site analysis and identification through comparison
against a database via wireless communication.
Such major improvements in the technology will take several years at least to
perfect. It is generally believed that STR markers will remain the standard for DNA
identification over the next decade, with mtDNA increasingly used to analyze highly
degraded samples. Research to develop ways to amplify longer sequences may,
however, lead to more use of VNTR or other more variable systems in the near
future. Also, the Human Genome Project and related efforts are increasingly
identifying very small genetic differences, even at the level of a single base (single
nucleotide polymorphisms, or SNPs). Such research may lead to eventual
development of better marker systems. In some cases, relevant physical
characteristics, such as for eye color, might even be deducible from genetic sequences
and could help, for example, in identifying suspects.59
58 See Amy Argetsinger, “Court Opens Door to New Paternity Challenges,” Washington Post,
Thursday, 29 June 2000, sec. A, 1, 19.
59 These and other potential developments are being considered by the DNA Commission and
There are several challenges raised by the likely future improvements in DNA
typing. Improvements in forensic typing will need to be validated by the scientific and
law enforcement communities and accepted by the courts before they become widely
available for use in cases. Therefore, any given case is likely to involve technologies
that are technically sound but not the most recently developed. Furthermore,
significant pressures exist to maintain a substantial degree of stability in the systems
used. Not only is adopting a new technology costly, but frequent change can lead to
longer and more costly proceedings and to uncertainty about the most appropriate
approach.60 One concern raised is that technological advances could lead to a
lengthening of the appeals process in cases involving DNA evidence, or a flood of
postconviction petitions for typing or retyping in situations where it would not
actually be helpful. Such concerns will need to be balanced against whatever added
ability the advances provide in reversing or avoiding wrongful convictions.
As the power of DNA profiling to identify a person continues to improve, the
question arises, can a person be positively or uniquely identified from a DNA profile?
FBI experts will currently testify that a person is the source of a DNA sample,
provided that the match probability is low enough (see the section on interpretation
of DNA evidence above). However, it has not yet been settled generally what
probability, or alternatively how many loci, are needed to effectively eliminate the
possibility that a match could be coincidental. One complication is that match
probabilities do not take into account the possibility of misidentification resulting from
errors at the laboratory or in the evidence custody chain. Both experts and the courts61
have generally agreed that such possibilities are best addressed in other ways.
The stability of DNA evidence makes it potentially useful not only in
postconviction typing; it also raises the question of whether statutes of limitations
should be extended for crimes in which such evidence is potentially important. That
question will likely increase in importance as the number of DNA profiles from “cold
cases” grows in databases and indexes such as CODIS. The benefits of using DNA
evidence to bring effective prosecutions after a longer period than currently feasible
will need to be balanced against factors such as the desire to avoid retraumatizing
victims who have recovered from the effects of crimes and the need to prioritize the
use of limited law-enforcement resources.62
are described on the commission’s website at [http://www.ojp.usdoj.gov/nij/dna/
60 For example, when the first NRC report on DNA evidence proposed a new method of
calculating match probabilities (the ceiling principle — NRC, DNA Technology, 82–85),
considerable controversy was generated in the courts and was not settled until the second NRC
report was released (NRC, Evaluation of DNA Evidence).
61 Those approaches include proficiency testing, laboratory accreditation, and providing
opportunities for retyping of evidence by the defense. For a discussion, see NRC, Evaluation
of DNA Evidence, 85–87, 179–185.
62 For discussion of this and other legislative issues, see Fischer, DNA Evidence.
Relationship of DNA Identification to Medical Genetic Testing
Both similarities and differences exist between the use of DNA for identification
and for medical genetic tests — with respect to techniques, applications, and the
issues raised. The most fundamental similarity is that both rely on genetic differences
among people. The most fundamental difference is in the characteristics of the DNA
sequences that each currently uses — in particular, variability, functionality, and
Variability. A genetic marker used in identification should be highly variable —
that is, there should be many alleles, and none of them should be very common. The
more variable the markers, the fewer are needed for a positive identification. In
contrast, a gene examined in a genetic test is unlikely to be highly variable. That is
because genetic tests are most often used in medicine, to diagnose or predict the
likelihood of a disease or condition caused by a genetic abnormality, which most
people will not have.63
Independence refers to whether different loci tend to be inherited together. Two
loci that occur close together on a chromosome will usually tend to be inherited
together — they will be linked, not independent. Two loci that occur on different
chromosomes will usually be inherited independently — they will be unlinked. In
genetic testing, only one locus is usually of interest, so linkage is not usually
important. In contrast, use of DNA to identify people requires examining several loci,
and the mathematics that is used works best if the loci are independent.64
Functionality. A sequence is of interest in genetic testing specifically because
it is a gene that codes for a particular chemical product. However, in identification,
a noncoding sequence or marker is of most interest.65 That is because the
mathematics that is used in identification works best with noncoding loci.66 Although
most loci used in identification are noncoding, it is possible that functions for at least
63 There are other potential uses of genetic tests, such as to predict how a patient is likely to
respond to a particular drug, but variability at any given locus will usually be low in such
cases as well. Also, genetic tests often use techniques other than direct identification of
sequences — such as visual examination of chromosomes (karyotypes), or examination of
64 In the simplest case, the frequencies of a particular genotype for each locus are simply
multiplied together to get an overall likelihood (see section on calculation above). While most
medical genetic tests do not currently involve more than one locus, there will likely be an
increased focus on multilocus testing as medical genetic technology becomes more
sophisticated. However, it is only for identification that independence of the loci is an
65 See section on why DNA can be used in identification at the beginning of this report.
66 For nonfunctional loci the effects of natural selection, which are difficult to quantify, do not
need to be taken into account when calculating the likelihood that someone will have a
particular combination of alleles; the probability can be calculated directly from the
frequencies of those alleles in a population (if certain other assumptions apply). For example,2
in the simplest case, if the frequency, x, of a particular allele is 10%, then x, or 1%, of the
population will have only that allele.
some will be discovered in the future. Furthermore, some, in particular STRs, are
thought to be linked to coding loci that may be implicated in genetic diseases or
Regulation and accreditation of laboratories performing medical genetic tests and
DNA identification are accomplished through different mechanisms. Clinical
laboratories, which may perform medical genetic tests, are regulated and must be
certified by the U.S. Department of Health and Human Services through the
provisions of the Clinical Laboratory Improvement Amendments of 1988, as amended
(P.L. 100-578). Standards are developed by the Health Care Finance Administration
and the Centers for Disease Control and Prevention. Medical research laboratories,
if they are supported by federal funds or are otherwise subject to regulation, are
subject to federal policies relating to the protection of human subjects (45 C.F.R. 46,
21 C.F.R. 50 and 56). Forensic laboratories do not require federal certification, but
to be eligible for DNA Identification Grants or Byrne Grants, they must adhere to
standards set by the FBI and undergo regular proficiency testing (P.L. 103-322, Sec.
210302). Also, in criminal and civil cases, DNA evidence must pass the scrutiny of
the judicial process, where custody-chain and quality-control procedures, as well as
other aspects of typing, may be challenged.68 Consequently, new DNA technologies
are likely to be adopted more slowly for forensic use than for medical testing.
Privacy and Discrimination
There is currently no federal law governing genetic privacy and discrimination
per se, although Congress has considered several bills addressing such concerns from
a medical perspective. The executive branch has also taken some actions, as have69
several states. The Privacy Act of 1974 (5 U.S.C. 552a) places restrictions on
agencies with respect to the disclosure of personally identifiable information in their
possession, including any “identifying particular assigned to the individual, such as a
finger or voice print or a photograph.” However, it applies only to federal agencies.
Executive Order 13145 prohibits genetic discrimination against executive branch
employees; it applies to medical genetic tests and does not cover the use of DNA in
identification per se. Other federal laws and guidelines provide privacy protections
67 One, VWA, is actually a noncoding segment of the human von Willebrand factor gene,
which is associated with a blood condition (James F. Crow, “Research and Development
Working Group Report,” Proceedings, National Commission on the Future of DNA
Evidence, 27 September 1999, [http://register.aspensys.com/nij/dnamtgtrans7/trans-k.html]).
68 For a discussion of those and other legal issues and cases, see NRC, Evaluation of DNA
69 For discussion of legislative issues and federal and state activities relating to genetic
discrimination and medical records privacy, see Nancy Lee Jones, Genetic Information: Legal
Issues Relating to Discrimination and Privacy, CRS Report RL30006, 2 October 2000; and
C. Stephen Redhead, Harold C. Relyea, and Gina Marie Stevens, Medical Records
Confidentiality, CRS Issue Brief IB98002, 2 October 2000.
for certain other kinds of personal information relating to health and medical
Current federal and state laws also provide safeguards against the misuse of
DNA profiles collected in law enforcement. Specifically, the DNA Identification Act
of 1994 (P.L. 103-322) required grantees to certify that DNA samples and analyses
“be made available only —
(A) to criminal justice agencies for law enforcement identification purposes;
(B) in judicial proceedings, if otherwise admissible pursuant to applicable statutes
(C) for criminal defense purposes, to a defendant, who shall have access to
samples and analyses performed in connection with the case in which the defendant
is charged; or
(D) if personally identifiable information is removed, for a population statistics
database, for identification research and protocol development purposes, or for
quality control purposes….
The act also applies those privacy requirements to federal, state, and local
participants in CODIS. However, state laws vary with respect to how such samples71
can be used. Some disagreement exists over whether state protections are adequate.
Samples deposited in the Armed Forces Repository of Specimen Samples for the
Identification of Remains are limited to the following uses: “Identification of human
remains; …[i]nternal quality assurance activities…; [a] purpose for which the donor…
(or surviving next-of-kin) provides consent;” or, when specifically authorized,
criminal procedures for which “[n]o reasonable alternative means for obtaining a
specimen for DNA profile analysis is available….” (DODD 5154.24, Sec. 3.5).
Since most genetic markers currently used in DNA identification are noncoding,
the privacy and discrimination issues raised are somewhat different than with medical
genetic testing. For example, knowing someone’s DNA profile for the 13 core STR
loci can provide information about paternity but provides no information at this time
about physical traits such as a propensity to a particular genetic disease.72 However,
the issues may begin to converge as new advances in genomics are developed and
applied, or if current identification technologies are applied in the context of genetic
testing. For example, in some cases clinical research laboratories have reportedly
used STR typing of samples as an added protection against the possibility of potential
70 See Redhead, Medical Records Confidentiality, 4–8.
71 For two different views, see Adams, “Hearing Statement,” and Barry Steinhardt, Statement,
Legislative Hearing on H.R. 2810, the “Violent Offender DNA Identification Act of 1999",
H.R. 3087, the “DNA Backlog Elimination Act”, and H.R. 3375, the “Convicted Offender
DNA Index Systems Support Act”; Subcommittee on Crime, House Committee on the
Judiciary, 13 March 2000, [http://www.house.gov/judiciary/stei0323.htm].
72 It is, however, possible that some relationships will be discovered in the future, as the
functions of human DNA sequences become better understood.
More generally, as DNA identification becomes more sophisticated, its potential
applications are likely to broaden, potentially increasing privacy concerns. For
example, it is already possible for parents to purchase kits for obtaining DNA samples
from their children and to have profiles developed from those samples and placed in
a database to be used potentially if the children become missing. Such private-sector
practices are not covered by federal laws and judicial rulings that provide privacy or
other protections against misuse or abuse of the information.
Government agencies in other countries have also become involved in DNA-
identification activities that would likely be limited to the private sector in the United
States. Britain’s Forensic Science Service (FSS) is an executive agency of the British
Government but is permitted to operate as a nonprofit corporation that offers services
not only to the government but also to the public. Among its services to corporations
is the development of DNA profiles of employees who might be at risk of being
kidnapped. The profiles could then be compared to any biological material provided73
by kidnappers. FSS also provides paternity-analysis services to the public. The
privacy of genetic information obtained or used in such services would presumably
be protected by a European Union directive on privacy of personal data.74
Such broadening applications may raise concerns about “function creep,”75 the
gradual widening of an application to uses not originally intended. One example often
given is the slowly broadening range of uses over the years of Social Security
numbers. The identifying power of DNA, the decreasing cost of typing, and the
increasing ability to obtain a useful sample without drawing blood, make DNA
potentially attractive for a wide range of uses requiring verification of identity.76
As long as DNA typing uses noncoding loci, privacy issues arising from its use
in identification should remain limited. However, three potential issues might deserve
particular attention as the use of the technology increases. First, as genomic research
leads to increasingly sophisticated technologies for detecting genetic differences, it
may become possible to use coding loci (genes) to provide identification. Use of
coding DNA for identification could raise issues of privacy and discrimination similar
to those that have raised concerns with respect to medical testing — if, for example,
such information became publicly available during a judicial proceeding.
Second, DNA samples obtained from a person — and in many cases from crime-
scene evidence, remains, or other sources — contain the person’s entire genetic code,
73 See the FSS website, [http://www.forensic.gov.uk/forensic/entry.htm].
74 Directive 95/46/EC of the European Parliament and of the Council of 24 October 1995 on
the protection of individuals with regard to the processing of personal data and on the free
movement of such data, available at [http://europa.eu.int/eur-lex/en/lif/dat/1995/
75 Steinhardt, Hearing Statement.
76 For a general discussion of such issues for biometric technologies, of which DNA typing
is one example, see Congressional Research Service, Biometric Science and Technology for
Personal Identification: Devices, Uses, Organizations, and Congressional Interest,” by
William C. Boesman, CRS Report RL30084, 8 March 1999.
not just the profile information. Therefore, the disposition of the samples themselves,
after profiling, is potentially an issue. That is especially a potential concern with
respect to private-sector activities where disposition of profiles or samples is not
necessarily regulated by current law. A case involving electronic commerce illustrates
the concern. Toysmart.com collected information about customers under a privacy
policy that claimed the information would not be shared with third parties. However,
when the company filed for bankruptcy, it included the database of customer
information among the assets it was selling. The Federal Trade Commission
unsuccessfully opposed the sale of the database.77
Third, a DNA sample and profile contain information not only about the subject,
but also about that person’s biological relatives. Therefore, consideration of privacy
issues related to DNA identification, as with genetic testing in medicine, must take
into account potential impacts on family members. That can raise potentially difficult
issues, as illustrated by the following hypothetical example. Suppose that a DNA
profile from a convicted offender is similar but not identical to that obtained from a
crime scene. Is it appropriate based on that information for a sibling of the convicted
offender to be arrested as a suspect? Alternatively, suppose that a person is suspected
of committing a crime but there is insufficient evidence to make an arrest. Suppose
further that the person has a sibling who has a profile in CODIS. Is it appropriate to
examine the profile of the sibling to determine how similar it is to the crime-scene
evidence? Such examples are likely to arise in real cases as DNA becomes more
widely used in identification.78
The growing use of DNA as an effective identification tool, and its increasing
overlap with aspects of medical genetic testing, are likely to create a range of policy
challenges over the next several years. While this report has discussed several current
and emerging issues, new ones may well develop as the technology evolves. The
biological role of DNA, the information it contains about family members, and other
features of the molecule and the technology may make some of those issues, and the
appropriate legislative response, especially challenging.
77 Matt Richtel, “FTC Moves to Halt Sale of Database at Toysmart,” New York Times, 11
July 2000, sec. C, 2; “Judge Shelves Plan for Sale of Online Customer Database,” New York
Times, 18 August 2000, sec. C, 2.
78 See Michelle Hibbert, “DNA Databanks: Law Enforcement’s Greatest Surveillance Tool?”
Wake Forest Law Review 34 (1999): 782–787 for a discussion of this issue and an application
to an actual case.