mcveanlab / mccortex

De novo genome assembly and multisample variant calling
https://github.com/mcveanlab/mccortex/wiki
MIT License
113 stars 25 forks source link

Tackle too much variation in a region #66

Open dimitrisarnellos opened 6 years ago

dimitrisarnellos commented 6 years ago

Hi,

I have an issue similar to this: https://github.com/mcveanlab/mccortex/issues/55

I am trying two different mapping strategies. On the first one, there are 4 flanks and 2 of them map, resulting in a variant to be called. On the second one, all of the 4 flanks map. There are two bubbles and the bubble that is being defined by the 2 flanks from the 1st mapping strategy, is being found nested in a bigger bubble on the 2nd mapping strategy.

On the second time, there are too many variants, and although the variant from the 1st time is being found, it gets this output from vcfcov: NC_000962.3 916 . A G . PASS BUBBLE=0;K21 K21R:K21A .:.

So it is not being called. It obviously is present since it is found the first time, but when I managed to map all the flanks, I loose this variant.

As has been mentioned on the linked issue 55, it most probably is due to too much variation in the region, but how does this work and is there a way to circumvent this?

Thank you in advance, Dimitris