SubstitutionSafari – from substitution calls to substitution spectra

Using the functions in SubstitutionSafari, we generate substitution – also termed single-base substitution (SBS) or single-nucleotide variant (SNV) – spectra in line with the widely-used classification scheme first adapted by Alexandrov et al., 2013 (see COSMIC SBS signature catalogues).

Note that, in order to produce your own spectrum, you will need to specify:

• the reference genome fasta file based on which your alignments' substitution calls were generated
• a VCF or dataframe object containing one substitution per line, featuring the minimal set of columns "CHROM", "POS", "REF", "ALT"
• a table with the triplet frequency counts of your reference genome or exome, which can be used to normalise for context-specific mutation opportunity (an example table can be found here)

My code then groups all of your substitutions with reference to the 96 different, pyrimidine-centered subgroups of C>A, C>G, C>T, T>A, T>C, and T>G types. Plotted spectra look like this example:

In the above case, we see enrichments of C>T substitutions, particularly in A[C]G and G[C]G contexts. These spikes indicate the presence of COSMIC SBS6, a mutational signature commonly identified in tumours with DNA mismatch repair deficiency. You might want to corroborate this by inspecting and fitting combinations of COSMIC signatures to the spectrum, for example by using a tool such as sigfit.

This script has been extensively benchmarked with substitution calls made against the Tasmanian devil reference. Though in theory, the code should run smoothly for ANY species with a reference genome.

Find it helpful or you simply enjoy making SubstitutionSafari spectra against your brand-new, awesome reference genome? Then why not surf the wave with your indel calls? – have a look at its sister library Indelwald! If you do face a challenge in using the code or wish to provide general feedback, please get in touch directly via [email protected]

---

Citation

If you can make good use of these functions in your work, I would be grateful for your citation of our associated preprint: The evolution of two transmissible cancers in Tasmanian devils (Stammnitz et al. 2023, Science 380:6642)