Volume 6, Issue 1 , Pages e8-e10, January 2012
The Israel Police DNA database: Recognizing the capability of databases in providing investigative leads
Article Outline
Dear Editor,
The Israel Police DNA database has been operative since February 2007. During this period of operation a number of criminal cases have been encountered and resolved not by direct cold hits, but by following up investigative leads provided by the DNA database. DNA databases contain a wealth of investigative information not yet routinely exploited. Partial matches obtained by DNA database searches, although presently not universally accepted, are rapidly becoming significant sources of investigative assistance for law enforcement agencies. In unsolved cases where all other investigative leads have been exhausted, a profile containing a large amount of allelic similarities to that from a crime scene may point the detectives toward a sibling or close relative who may be the actual perpetrator of the crime. In a number of global databases, law enforcement requests for such searches in the form of “familial searches” are becoming commonplace. In this letter we present a number of cases and scenarios encountered during the routine operation of the Israel DNA database illustrating the benefit that can be afforded to the local investigative process when databases are recognized and exploited for more than just as a source of direct hits.
Recently, a number of publicized cases solved by information generated from partial matches and familial searches, as in the case of California's Grim Sleeper [1], have shed light on the ability of DNA databases to provide intelligence leads which supply valuable investigative information toward identifying unknown perpetrators of crime [2]. Familial searching is defined as the deliberate searching of a DNA database to identify profiles that match at only a limited number of known loci, in an attempt to locate previously unknown relatives in order to generate new investigative leads [3]. Partial searching is not specifically prohibited by legislation in the United States, and while the practice can provide powerful investigative aids, it is not yet widely accepted or utilized. On the other hand, the states of California and Colorado have already developed guidelines and policy for initiating such searches and reporting DNA partial matches in unsolved cases involving single-source profiles [4], [5], [6].
Using close matches to provide investigative leads is far from novel in DNA forensics. Even before the advent of DNA databases, highly similar profiles noticed by analysts in casework could provide leads for investigators to ultimately apprehend the actual perpetrator of a crime [7], [8]. As the recognition regarding the potency of these search methods grows, more and more specific requests to perform such “familial search” or partial matches on databases are being requested. Familial searching may be used to identify similar, but not identical DNA profiles within a database. DNA profiles that share a notable number of alleles may originate from related individuals and therefore performing a familial search within the database attempts to discover the identity of an individual who could possibly be the sibling, parent or offspring of the actual person being sought. This investigative technique has most often been employed in the United Kingdom to date and examples of success stories are abundant [9], [10]. In addition, partial profiles found on databases, resulting from failure to successfully amplify complete profiles or by use of complementary kits such as Minifiler™, may also be a source of information in providing leads.
These practices of partial matching done in an attempt to possibly identify a close relative of the potential suspect are raising debate. Critics fear that wide spread use of such searches may intrude on personal privacy issues and cast suspicion far too widely. It is feared that in this climate of DNA database expansion, many DNA profiles found on databases will be used to generate lists of possible suspects for committed crimes. In other words, essentially, people who are not considered to be directly involved in specific crimes will become “suspects” and investigated just by means of having their genetic information found on a database. In areas where there are large ethnic groups and higher circumstances of inbreeding, this can become even more worrisome.
Products of partial searches do not automatically provide viable leads. Specifically searching for partial matches often generates long lists of possible candidates, depending on the frequency of specific pairs of alleles found at each target locus in the profile searched. The list must then be reduced to produce a list of more probable candidates. This is accomplished by conducting a number of successive tasks, including the calculation of likelihood ratios (LRs) where the plausibility of a parent–child or sibling relationship is measured numerically. Higher LRs give more credence to the assumption that two profiles originate from related individuals [11]. If the profile in question originates from a male origin, the list will further be refined by carrying out Y-STR haplotype analysis on both the crime scene item and the list of database leads, which will eliminate the majority of those not related by male lineage. This exercise is only possible in countries where legal restrictions do not prohibit the storing of reference samples for future confirmatory testing.
The Israel Police DNA database was established according to legislation passed by the Israel Knesset in 2005 and has been operative since February 2007. To date, Israel's database contains over 100,000 profiles from suspects, convicts and arrestees and more than 6000 forensic profiles. The workflow used for the database was developed and validated in-house where the sample of choice is buccal cells collected on FTA cards. The entire process from sample registration through profile analysis is directed by an in-house built LIMS where ultimately, profiles generated from the reference samples are automatically uploaded to the database. Crime scene profiles from items of evidence are electronically transferred daily directly from the casework lab files onto the database containing the offender profiles. Each new profile uploaded is automatically searched against all existing offender and forensic profiles within the database and if a match occurs, the samples are retested to confirm, and a hit report is generated. According to our legislation, one of the primary goals of the DNA database is to assist in crime investigations. Although there is no specific stipulation regarding familial searching within the Israeli legislation, it is not the practice of the Israel DNA database to deliberately carry out familial searches.
The computer software program employed and developed by the Israel database is a program containing a built-in capacity to not only search for completely matched hits, but also to identify similar profiles where a few mismatched loci may occur within a profile. Originally this utility was added to the program to take into account possible human error when crime scene profiles from the casework lab were manually uploaded to the repository. Therefore, in addition to a full match search, profiles that are identical at only 7 out of 11 loci are also noted. This utility also is able to identify matches between incomplete and full profiles. Our database laboratory guidelines permit incomplete reference profiles of no less than seven of the eleven SGM Plus™ STR loci to be uploaded to the database. Forensic sample profiles derived by the casework lab of no less than eight of the 11 SGM Plus™ loci can be deposited on the database.
As a result of this “similarity search” routinely carried out with each upload, a lengthy list of similar profiles differing only at a small number of loci may be generated, depending on the frequency of alleles found within the profile. An integral output of our search process is a likelihood ratio (LR) calculation for siblings, which provides an indication of the strength of each of these partial matches. The accepted practice for the Israel database is to report to the casework lab only those partial matches with a LR of >200,000 and only those associated with violent crimes including rape, murder and aggravated assault. If the forensic unknown profiles originate from a male source, then these partial matches may be followed up by carrying out Y-STR typing on both the crime scene sample and the database sample to determine if they have identical Y-haplotypes. If an identical Y-haplotype is defined, this lends weight to the hypothesis that the two partially matched samples originated from subjects with the same paternal lineage.
We present here a variety of scenarios encountered since the commencement of operation of the Israel DNA database, examples illustrating where meaningful information inherently connected with existing samples in the database was further investigated and ultimately provided the necessary leads to resolve each case.
The first case presented is from an illegal arms smuggling case, a crime particularly prevalent in our region. An unknown casework profile from a sock used to package a grenade was uploaded to the DNA database. The search results provided a partial match (7 full matching loci/17 out of 20 identical alleles) to an existing known suspect profile in the database. A likelihood ratio calculation of probable siblings provided a convincing 11.7 million result. Y-STR results from both the evidence and the sample from the database provided an identical Y haplotype supporting the supposition that the profile actually did originate from a brother or another male from the same paternal line. Upon follow up inquiry by the investigating team, a sample was obtained from a “new” suspect, a brother of the database subject, and a match was reported. Numerous such partial matches are routinely noted and may be reported to the casework lab by the database lab and whether these leads can then be followed up is dictated to by such considerations as analyst workload and severity of the crime committed. In this case, the tip was followed up and aided in solving the crime.
The next example comes from a property crime casework sample involving exceptionally inhibited DNA. A blood stain was collected from a crime scene and attempts to obtain a suitable SGM Plus™ profile suitable for comparison to the database failed. A Minifiler™ profile was successfully generated and in order to compare it to the database, which contains profiles derived using the SGM Plus™ kit, a less stringent search was performed, specifically for the 5 shared loci between the two kits. A match to an existing known suspect profile in the database resulted at all five identical loci (a 10 out of 10 allele match). In order to verify this match, the suspect's sample was re-amplified, this time using the Minifiler™ kit. This 8 loci match provided a statistical frequency of over 26 billion within the Israeli population. In this case, only one sample from the database matched, but occasionally a list of a number of possible candidates is created, and then can be further reduced by examining such additional demographic parameters associated with the database such as residence of the subject, location of crime committed, etc. In these cases further Y-STR testing may also be done to support the association.
Our last example comes from a sexual assault case that resulted in a pregnancy which was consequently aborted. The perpetrator of the assault and alleged father of the fetus was unknown. Tissue from the aborted male fetus and the mother were genotyped using the SGM Plus™ kit. As an investigative tool, the Police DNA database was searched for the identified obligatory paternal alleles from the fetus. The search resulted in the identification of two suspects found on the database who could be the possible father of the fetus according to the seven obligatory alleles defined. Subsequent genotyping was performed using the Y-Filer™ kit which resulted in the exclusion of one of the two suspects. PowerPlex 16™ typing was then carried out on the remaining suspect's sample and evidence sample in order to strengthen the statistics of the match by adding additional loci. As a result of the new information acquired from these supplementary tests, the victim was re-questioned and admitted to having an affair with the newly identified suspect.
Without entering the debate regarding the controversy over the use of database searches to identify relatives of an unknown suspect, we present here cases where additional investigation and clarification of partial information associated with samples found on our database lead to the identification of actual perpetrators in a number of cases. Such scenarios are routinely encountered in DNA databases worldwide.
These examples demonstrate how data found within databases can potentially be more informative than solely being used to provide direct hits. As a result of the enormous workloads and backlogs already present in many laboratories, these steps may not routinely be carried out expediently and therefore, possible leads from partial matches may not be exploited.
If the ultimate goal of all forensic laboratories is to provide scientific evidence toward solving crimes, then we suggest that the information obtained from partial matches should also be regarded as a legitimate resource and an intelligence tool. Just as the discovery of a DNA profile at a crime scene does not automatically prove the source's guilt, so should leads provided by confirmed partial matches be regarded.
References
- http://articles.latimes.com/2010/jul/10/local/la-me-0710-grim-sleeper-dna-20100710.
- . Turning base hits into earned runs: improving the effectiveness of forensic DNA data bank programs. J. Law Med. Ethics. 2006;34(2):222–233(12)
- Forensic utilization of familial searches in DNA databases. Forensic Sci. Int. Genet. 2010;
- Information Bulletin, Bureau of Forensic Services, California Department of Justice, Division of Law Enforcement. No 2008-BFS-01.
- Searching for first-degree familial relationships in California's offender DNA database: validation of a likelihood ratio-based approach. Forensic Sci. Int. Genet. 2010;
- M. Morrissey, DNA-crime scene to courtroom, Presented at the National Institute of Justice Post Conviction DNA Case Management Symposium, Palm Harbor, FL, 2009. Available from: http://projects.nfstc.org/postconviction/day1.htm.
- . Finding criminals through DNA of their relatives. Science. 2006;312:1315–1316published online 11 May 2006; doi:10.1126/science.1122655
- . The story of Abraham, Isaac and Jacob or am I my brother's keeper?. J. Forensic Sci. 2003;48(1):137–139
- . Effectiveness of familial searches. J. Forensic Sci. Soc.: Sci. Just. 2008;48(4):164–167
- Familial DNA Database Searches, http://www.denverda.org/DNA/Familial_DNA_Database_Searches.htm (last visited 16 April 2010).
- . Forensic DNA Typing. 2nd ed.. New York: Elsevier; 2005;pp. 510–513
PII: S1872-4973(10)00189-4
doi:10.1016/j.fsigen.2010.11.006
© 2010 Elsevier Ireland Ltd. All rights reserved.
Volume 6, Issue 1 , Pages e8-e10, January 2012
