However, the newly developed approach for deciphering mutational signatures also allows extending mutational signature analysis over an arbitrary selected set of biologically meaningful mutation types

[20••]. To demonstrate its applicability, the mutational catalogues of the 21 breast cancer genomes were extended to include double nucleotide substitutions, indels at microhomologies, indels AZD4547 cell line at mono/polynucleotide repeats, and even a complex mutation type such as kataegis. Reanalysing these mutational catalogues demonstrated that kataegis separates as its own mutational process. Further, double nucleotide substitutions and indels at microhomologies associated predominantly with the activity of the previously identified uniform mutational process. Lastly, indels at mono/polynucleotide repeats did not strongly associate with any of the previously described mutational processes [ 20••]. Extending the previously defined mutational catalogues illustrated the possibility of incorporating additional mutation types and it revealed some associations between substitutions Selleck GSK2118436 and indels thus providing more biological insight into the identified mutational processes [20••]. Further biological insight was derived by analysing mutational catalogues that incorporate the transcriptional strand on which a substitution resides in the footprints of a gene. Thus, the previously

defined 96 substitution types were extended to 192 mutation types. For example, the number of C > T mutations at TpCpA were split into two categories: the number of C > T mutations at TpCpA occurring on the untranscribed strand of a gene and the number of C > T mutations at TpCpA occurring on the transcribed strand. In general,

one would expect that these two numbers are approximately the same unless the mutational learn more processes are influenced by activity of the transcriptional machinery. This could happen, for example, due to recruitment of the transcription-coupled component of nucleotide excision repair (NER) [87•]. If a mutational process has a higher number of C > A substitutions on the transcribed strand compared to the C > A substitutions on the untranscribed strand (i.e. note that C > A mutations on the untranscribed strand is the same as G > T mutations on the transcribed strand), this could indicate that the mutations caused by this process are being repaired by NER. As such, this analysis provides a further insight into the operative mutational processes and their interaction with cellular repair processes. A known example of such strand bias due to interplay between a mutational process and a repair mechanism is the formation of photodimers due to UV-light exposure that are repaired by NER and result in a higher number of C > T mutations on the untranscribed strand [87•].