create a homopolymer score to weight the probability of a given variant call

[aineniamh]

i think actually just use the q score given by medaka once the pipeline is updated

[aineniamh]

have been testing this, but doesn't look like a good metric for variant quality