Search

Cigarette smoke comprises a large number of carcinogenic substances that can increase DNA mutation load in epithelial cells of the mouth, throat and lungs. While a strong C:A substitution preference is abundant in tobacco-related cancer genomes, detection of complex or less abundant somatic mutation signatures may be confounded by the heterogeneity of carcinogens present in smoke. Trinucleotide signatures are defined for a variety of somatic mutation processes, yet the extent to which this configuration optimally defines and discriminates between mutational processes is not clear. Here, we describe a method that determines whether trinucleotide patterns do a good job at encapsulating a mutation signature or whether they mask underlying heterogeneity that alternative pattern structures would better define. The approach works by mapping the dependency of trinucleotide signatures in relation to sequence context to establish a 'footprint' of context dependency. Applying this technique to smoke-associated cancers, we show that a robust tetranucleotide substitution is prevalent in 17% of lung squamous cell carcinoma genomes. The signature is dominated by the substitution CT(C:A)G and is strongly associated with gene expression level and intron-exon junctions. Intriguingly, its distribution across the genome is biased towards 5' splice junctions, suggesting a novel mechanism of mutation. [ABSTRACT FROM AUTHOR]

Summary: QuasR is a package for the integrated analysis of high-throughput sequencing data in R, covering all steps from read preprocessing, alignment and quality control to quantification. QuasR supports different experiment types (including RNA-seq, ChIP-seq and Bis-seq) and analysis variants (e.g. paired-end, stranded, spliced and allele-specific), and is integrated in Bioconductor so that its output can be directly processed for statistical analysis and visualization. [ABSTRACT FROM AUTHOR]