Forensic inference from genetic markers uses highly polymorphic multi-locus genotypes. Measures of informativeness can aid in selecting efficient genetic markers. Existing measures do not account for multiple sources of genetic variation (i.e. mutation, silent alleles, etc.) and they are not directly applicable to complex identification problems. Using information theoretic principles within a probabilistic expert system (PES) we define a general measure of informativeness, I_q, of a marker for answering a forensic query. I_q gives a slightly different ranking of most genetic markers as its comparable measures. Accounting for sources of variation such as mutation, silent and null alleles reduces I_q and may further affect ranking. This criterion has a solid theoretical basis and can account for multiple sources of genetic variation and other anomalies. It can be directly applied to a variety of planning issues concerning the type, quantity and specific choice of markers for use in paternity testing and more general forensic problems.