Figure 1.

The mutation spectra and mutation rate of lung cancer genomes. (A) Lung cancer genomes have a few distinct patterns of mutation composition. Genomes from smokers tend to have a large fraction of C:G > A:T transversions, a known tobacco-related DNA-damage pattern (top panel), and have larger number of filtered variations (bottom panel). In addition, the four cell lines with the most protein-altering SNVs all have mutations in at least one of the mismatch repair (MMR) genes. (B) CDS and UTRs have the lowest mutation rates among genomic features. Mutation rates (i.e., number of mutations per mega base) for different genomic features were calculated, and then normalized by the genome-wide mutation rates to obtain the relative mutation rates. Each dot in the plot represents one genome (blue, cell lines; red, tumors). The boxes in the box-and-whisker plots represent the interquartile range between the first and third quartiles; the dashed lines (whiskers) extend to the most extreme data points which is no more than 1.5 times the interquartile range from the box. (C) Mutation rates are negatively correlated with expression level. In each genome, genes were divided into three groups according to their expression level based on the RNA-seq data from the same sample: zero (RPKM = 0), low (0 < RPKM ≤ 1), and high (RPKM > 1). Mutation rate for each group was calculated, and then normalized by average mutation rate for all genes in the genome to obtain relative mutation rates. Genes with high expression levels tend to have the lowest mutation rates in either the entire intragenic region (i.e., exons and introns, left panel) or introns only (right panel). This suggests that transcription-coupled DNA-repair mechanisms may play a role. The boxes in the box-and-whisker plots represent the interquartile range between the first and third quartiles; the dashed lines (whiskers) extend to the most extreme data points which are no more than 1.5 times the interquartile range from the box.