Fig. 3

Results of targeted deep sequencing to compare genetic alterations between patient-derived and second-generation tissues (a) We used Venn diagrams to demonstrate the overlapping somatic mutations (single-nucleotide variants and insertions/deletions) for YHIM-3002 and YHIM-3009 samples. The Jaccard similarity score was used to measure the similarity; the score of most samples was > 82%. The highest scoring models were YHIM-3002 and -3009, with respective Jaccard similarity scores of 93.6 and 95.2% (Fig. 3a). The MAF values for common mutations between F0 and F2 demonstrated overall concordance (r² = 0.93 and 0.95 for YHIM-3002 and -3009, respectively), implicating that most germline mutations in the F2 sample corresponded with those in the F0 sample (Fig. 3b). Heatmaps showing mutational overview of known functionally active genes and significantly mutated genes in patient-derived xenografts (PDXs). Oncomine Cancer Panel was used to detect somatic mutations in PIK3CA, HRAS, and TP53, as well as EGFR, CCND1, MYC, and PIK3CA amplification. Heat maps showed all Oncomine-defined relevant alterations in the RNA (header) and DNA components of the 14 PDX specimens. Red indicates amplification, green indicates a missense mutation, orange indicates a small insertion/deletion, purple indicates a fusion, and black indicates a multi-hit result (Fig. 3c)