A new strategy for body source identification of tumor sample

Abstract Body source identification of tumor sample with likelihood ratio is usually difficult because of the high mutation rate of STR in tumor. Here, we developed a new strategy to identify the body source of tumor sample. In this study, 423 tumor and their homogenous normal samples were genotyped with Goldeneye 20A multiplex STR genotyping system, which was including 19 STR loci widely used in forensic casework. An improved identity by state (IBS) scoring system was adopted to calculate the IBS score, the number of loci with 2 alleles ( A 2 ), 1 allele ( A 1 ) or 0 allele ( A 0 ) sharing within each tumor-normal pair (TNP). The binomial distribution of the four variates in unrelated individual pairs (UIP), parent–offspring pairs (POP) and full-sibling pairs (FSP) were derived based on the allele frequencies of the 19 STR loci in Chinese Han. The results showed that 30.97% of tumor samples carried one or more STR loci with genotypic alteration. According to the binomial distribution of the four variates in the studied groups, A 2 was most effective to distinct TNP from UIP, TOP, or FSP. Using the criterion of A 2 ≥15, the body source of 98.82% of tumor samples was successfully determined. Contrarily, the probability of A 2 ≥15 in UIP, TOP and FSP was 0.00%, 0.00% and 0.02%, respectively. We further demonstrated that the strategy of A 2 binomial distribution was more powerful than likelihood ratio method in body source identification of tumor sample.