We identified a total of 32 somatic mutations across the four xenograft samples, of which 63% (43/68) recurred in more than one sample and corresponded to mouse-human mismatches. An investigation revealed that a fraction of the PCR primers amplified both human and mouse DNA and that these mutations corresponded to mismatches in the mouse-human genome multiple alignment (Figure S6 in Additional file 1). To filter out false positive mutations in the xenograft samples coming from mouse DNA contamination, we obtained the multiple-alignment Mouse-Human track from the UCSC Genome Browser. We then identified all mismatch positions between hg19 and mm9 genome sequences located in the UDT-Seq amplicons. We filtered out the somatic mutations in the xenograft samples, with coordinates corresponding to these mismatched positions. After removing these exogenous mutations, there were 1, 11, and 1 high confidence somatic mutations in the breast, colon, and ovarian xenograft samples, respectively. No significant somatic mutations were detected in the sarcoma xenograft sample (Table S7 in Additional file 2). To annotate the finalized list of filtered significant somatic mutations for functional consequences, we used ANNOVAR [30] on hg19.
