Add trinucleotide context visualisations

[selkamand]

If you have a lot of variants lacking RNA support despite a high DNA VAF and RNA depth, the mutations probably reflect library-related sequencing artifacts.

Since Sage annotates variants with their trinucleotide context, it may be useful to visualise the trinucleotide count matrix for these likely artificial variants. We should do the same for the real variants with RNA support (e.g. if you filter out the dodgy stuff)