Racial Classifications Based on Blood Group Protein Systems

Abstract

Two methods employing frequencies of blood group phenotypes are assessed as methods of accurately classifying individuals into racial categories. The data used consist of eleven blood group protein systems from 5043 males distributed into four ethnic categories (white, black, Chicano/Amerindian, and Asian). Both methods work equally well and yield a rate of misclassification of about 22%. Also included is a discussion of two probabilistic calculations relevant to employing blood group protein systems in the context of an identification tool.