Computational analysis of genetic paternity test

Abstract

The genetic paternity test aims to establish whether a certain individual, the claimant, is the child of a second individual, the respondent. The tests most commonly applied in the courts today use DNA (Deoxyribonucleic acid) to assess the probability of family relationship in disputed judicial paternity cases. Nowadays there is a vast literature on cases on which genetic material of the respondent is available. In the absence of this information, there are few references in the specialized literature. The doctoral thesis of Luiz Montoya-Delgado addressed this problem, but despite the technical advance it represented, specific cases were described for which specific techniques were presented. The objective of this work is to develop an automatic calculation tool which encompasses every type of judicial request and which will be efficient even on a large scale and with data obtained from individuals with different degrees of kinship. The main activities will be: 1) use of bayesian nets for modeling and calculation of the probability of paternity; 2) development of a heredogram- bayesian nets compiler; 3) computational and mathematical treatment of the mutations in micro- satellites (STRs-Short Tandem Repeats); and 4) hardy-weinberg equilibrium test, by means of a new methodology of tests of exact hypotheses. (AU)