When scientists use DNA evidence in court, coancestry effects such as popul
ation structure and relatedness are usually ignored. In paternity cases, on
ly if a particular man has the child's paternal allele at a certain locus,
can he not be excluded in the paternity dispute. However, it is certainly t
rue that close relatives will be far more likely to have the child's patern
al allele than will random members of the reference population. In particul
ar, the probability that the true father's brother has the paternal allele
is very much greater than that for any other relationship. In this paper, t
he authors describe a method for inference in a case where the true father
may be a relative of the alleged father. This paper also reports that most
current methods overstate the probability that the alleged father is the fa
ther.