Fig. 4

CNVs in different samples with boundary resolution at gene level. CNV gains and losses involving key driver genes in certain samples are shown here. Genome sequences are represented by the central axis, with genomic arm locations given on the right side. Their genomic starting and ending base pair locations were indicated in pink numbers ending with bp. Since our results are achieved through exome sequencing, the boundaries demarked here represent the contained regions within the real CNV blocks instead of the exact basepair locations. Gene IDs within the CNV blocks are either in yellow (a-c) or green (d) boxes, or listed above the central axis. Sample IDs are listed below the central axis, with their respective CN listed on the right side, in white. A CN greater than 2 is interpreted as CNV gains and a CN less than 2 as CNV losses. Sample IDs start with Patient IDs (P1-P5), followed bya hyphen and tissue type (S or L). S represents primary cancer tissue, and L metastatic lymph sites. a Likely two blocks of CNV gains involving ERBB2. One block starts with PP1R1B and ends with LRRC3C, with a genomic length of ~ 318Kb containing a total of 14 genes. Tissues include P2-L and P3-S. The other one, in P3-L, starts at GRB7 and ends afterLRRC3C. b Genomic block amplification in 17q21.1, starting at GSDMA and ending at gene WIPF2. The block size is ~ 319Kb, involving 4 tissues (P2-S, P2-L, P3-S, P3-L). This amplification is likely critical for both tumorigenesis and metastatics, as the amplification is present in both primary tumors and metastatic tissues with the exact same boundaries. CN is from 2.75 to 6.50. c Another block with CNV gains in 17q21.31. This block is the largest, with genomic length of 2,202Kb, and involving 63 genes. The boundaries are most likely fixed in three samples (P2-S, P3-S, P3-L). d A CNV loss event in 3 samples (P2-S, P4-S, P4-L) in chromosome 9p21.3, involving two cancer driver genes (CDKN2A and CDKN2B). The block size is 41Kb, a relatively small block. It is likely that one copy of these two genes was lost in one of the chromosomes. The differences in the CN in different tissues are small, and could simply be caused by the portion of cancerous cells within the sample tissues. This loss may be important to tumorigenesis and metastasis