Use of odds of paternity computations in determining the reliability of single exclusions in paternity testing

In parentage determination with genetic markers the concept of paternity index is used to evaluate the relative likelihood of fathership of an accused male individual. When a battery of genetic tests are performed, the event of inadvertant technical errors or the occurrence of suppressor genes that cause an apparent paternity exclusion of the male concerned reduces this likelihood to zero. However, through a study of statistical properties of the paternity index statistic based on the other genetic systems, it is possible to examine the reliability of one or a few such apparent exclusions. The details of this approach is examined here by analytical methods following the statistical principles of likelihood ratio test procedures. Some illustrative data are also analyzed to show the practical utility of this approach.