Abstract: Objective To establish and evaluate the method of matched locus numbers in full sibling identifica-tion.Methods Two hundred and eighty full sibling(FS) pairs and 2 003 unrelated individual(UI) pairs were genotyped with Identifiler system.The number of locus matched with 0 identical allele(A0),matched with 1 allele(A1) or matched with 2 alleles(A2) were counted and full sibling index(FSI) were calculated based on ITO method.Fisher discriminant functions were established based on the numbers of matched STR locus or FSI.Power of different Fisher discriminant functions was statistically analyzed.Results The distribution of A1 and A2 in FS group and that of A0 and A1 in UI group were in accord with normal distribution.Contrarily,A0 in FS group and A2 in UI group fitted to skew distribution,respectively.Differ-ence of A1 distribution was not statistically significant in the two groups(P>0.01).The established Fisher discriminant functions based on A0 and A2 for each group were ZFS=0.998 17 A0+4.244 42 A2-12.779 70 and ZUI=2.014 56 A0+1.546 58 A2-7.280 76,respectively,the average error probability of which was as low as 0.049 0.The power of discrimination for full sibling showed no statistically significant difference between ITO method and the established Fisher discriminant functions.Conclusion The number of matched STR locus in Iden-tifiler system is a valuable method in full sibling identification.The discriminating power of the estab-lished Fisher discriminant functions based on the matched STR locus number is similar with that of classic ITO method in full sibling identification.

Key words: forensic genetics, sibling relations, short tandem repeat, discriminant analysis, matched locus