Format

Send to

Choose Destination
Forensic Sci Int Genet. 2014 Nov;13:167-75. doi: 10.1016/j.fsigen.2014.07.015. Epub 2014 Aug 2.

Does an English appeal court ruling increase the risks of miscarriages of justice when complex DNA profiles are searched against the national DNA database?

Author information

1
Division of Forensic Sciences, Norwegian Institute of Public Health, Rikshospitalet, Sognsvannsveien 20, 0372 P.O. Box Nydalen, 0403 Oslo, Norway; Institute of Clinical Medicine, University of Oslo, Postboks 4950 Nydalen, 0424 Oslo, Norway. Electronic address: peterd.gill@gmail.com.
2
Division of Forensic Sciences, Norwegian Institute of Public Health, Rikshospitalet, Sognsvannsveien 20, 0372 P.O. Box Nydalen, 0403 Oslo, Norway.
3
Division of Forensic Sciences, Norwegian Institute of Public Health, Rikshospitalet, Sognsvannsveien 20, 0372 P.O. Box Nydalen, 0403 Oslo, Norway; IKBM, UMB, Postboks 5003, 1432 Ås, Norway.

Abstract

Likelihood ratio (LR) methods to interpret multi-contributor, low template, complex DNA mixtures are becoming standard practice. The next major development will be to introduce search engines based on the new methods to interrogate very large national DNA databases, such as those held by China, the USA and the UK. Here we describe a rapid method that was used to assign a LR to each individual member of database of 5 million genotypes which can be ranked in order. Previous authors have only considered database trawls in the context of binary match or non-match criteria. However, the concept of match/non-match no longer applies within the new paradigm introduced, since the distribution of resultant LRs is continuous for practical purposes. An English appeal court decision allows scientists to routinely report complex DNA profiles using nothing more than their subjective personal 'experience of casework' and 'observations' in order to apply an expression of the rarity of an evidential sample. This ruling must be considered in context of a recent high profile English case, where an individual was extracted from a database and wrongly accused of a serious crime. In this case the DNA evidence was used to negate the overwhelming exculpatory (non-DNA) evidence. Demonstrable confirmation bias, also known as the 'CSI-effect, seriously affected the investigation. The case demonstrated that in practice, databases could be used to select and prosecute an individual, simply because he ranked high in the list of possible matches. We have identified this phenomenon as a cognitive error which we term: 'the naïve investigator effect'. We take the opportunity to test the performance of database extraction strategies either by using a simple matching allele count (MAC) method or LR. The example heard by the appeal court is used as the exemplar case. It is demonstrated that the LR search-method offers substantial benefits compared to searches based on simple matching allele count (MAC) methods.

KEYWORDS:

Likelihood ratio; Matching allele count; National DNA database; Trawl

PMID:
25151459
DOI:
10.1016/j.fsigen.2014.07.015
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center