디엔에이(DNA)증거의 해석의 오류가능성과 증거법적 대안

Evidential and Procedural Measures against Possible Errors in the Interpretation of DNA Evidence

This article discusses evidential and procedural measures against DNA errors in criminal investigation and trials. It has been believed that a Random-match Probability(RMP) of DNA profile is so astronomical that a coincidental match in population is unlikely. RMP of DNA evidence, however, has been mistakenly considered as the probability of identifying criminals. In contrast, the RMP is merely frequency showing that two randomly selected persons might match in their DNA STRs, rather than a suspect likely being an offender. Therefore, with Bayes’ Theorem, Posterior Odds = Likelihood Ratio(LR, 1/RMP)× Prior Odds. Thus, if we don’t know Prior Odds, we cannot calculate Posterior Odds. As a result, if there is only LR evidence against suspects, they should not be convicted solely based on DNA profile random-match rate. Prior Odds, however, can be estimated with the size of population, even without further evidence. For e×ample, if a suspect lives in a community of 40,000 population, we can assume that the likelihood of the suspect being a culprit is 1/40,000. Then we can multiply 1/40,000 with LR. In this case, Posterior Odds(the likelihood that the suspect is an offender) can be calculated. There are some caveats, however. Other issues that raise the error rate of DNA evidence must be dealt with, such as chain of custody and pre-trial discovery. Chain of custody of DNA evidence must be observed and proved from the stage of DNA evidence gathering to a DNA e×amination process and e×pert testimony at trial. In addition, defendants must be given access to all the case files and data made at a laboratory and given an opportunity to rebut the reliability of test results prior to trials. Furthermore, indigent defendants must be assisted by a DNA e×pert funded by a government.