Problems in probability of paternity interpretation.

This paper discusses the statistical interpretation of blood group findings in paternity testing. As a consequence of the large number of systems now employed, high probabilities of paternity are usual and evaluation problems arise. The purpose of this investigation was to calculate the paternity probabilities for a sample of legitimate families with a true father compared with those obtained in some cases of non-excluded men chosen randomly from the population as the accused fathers for the same mother-child pairs. The calculations were based on Essen-Möller formula, derived from Bayes' Theorem. The blood group systems taken into account were ABO, Rh, MNSs, Kell-Cellano, P1, Duffy, Lutheran, Kidd, Gc, Hp, Gm, Km, Tf, alpha 1-AT, AcP, PGM1, AK, ADA, EsD, 6-PGDH and GLO-I. Applied together these give an exclusion probability of 97.32%. The results of probability of paternity for some mother-child-father triplets and its comparison with the chance of exclusion for the same mother-child pairs are reported.