As described earlier, the interpretation of DNA profiles obtained from mixtures is difficult at best. One especially dangerous warning sign is that many testing laboratories decline to draw conclusions regarding mixed samples in the absence of knowledge regarding the DNA profiles of individuals that are expected by investigators to be a contributor to a sample. The relevant question for mixed samples is “What fraction of the general population would be definitively excluded as being a possible contributor to this evidentiary sample?” It is possible to objectively address the possibility that alleles are masked by the presence of either alleles from other contributors or by technical artifacts (such as stutter peaks). Such approaches typically generate fairly unimpressive numbers – particularly when the discretion to dismiss a small number of “anomalous” results are taken into account as well. As a result, it is common (though not generally acceptable to the scientific community) for analysts to report the answer to a very different question, namely “What is the rarity of the reference sample in the general population?”
Kafarowski, E., Lyon, A., and Sloan, M. (1996). ``The retention and transfer of spermatozoa in clothing by machine washing.'' Canadian Society of Forensic Science Journal, 20(1):7.11.
Murphy, M.F. (2001). ``Review of Transfer Articles'' Memo reviewing several DNA transfer studies.
Perlin, M.W. and Szabady, B. (2001). ``Linear mixture analysis: a mathematical approach to resolving mixed DNA samples'' Journal of Forensic Sciences, 46(6):1372-1378.
Storey, J. (1998). DNAMIX 2. Software that uses likelihood ratios pertaining to mixed DNA samples.
Taylor, M. (2001). ``TA case #1458, Commonwealth v. Dirk K. Geineder.'' DNA lab report.
Wang, T., Xue, N., Rader, M., and Birdwell, J.D. (2001). Least Square Deconvolution (LSD) of STR/DNA mixtures. Association of Forensic DNA Analysts and Administrators Winter Meeting. January 17-18. Austin, TX.