Analysis of a mixed biological stain by means of highly polymorphic VNTR sy
stems usually reveals a profit composed of multiple markers. If the victim
and one or several suspects match the profile, the evidential strength of t
he matches has to be very carefully analysed. The appropriate methods for t
he statistical analysis of DNA profiles advanced recently were limited to c
ases with no relationship between the tested and non-tested persons. The pr
esent paper extends the theory beyond this limitation.