Adjacent SNV/Indel missing

[zhuchcn]

We intentionally ignore SNVs and Indels that are right next to each other. This was based on a trust on variant callers that it is called as a SNP/SNV than it must not be MNV. But turns out variant callers are not perfect and it does seem like adjacent SNPs can actually be MNVs in real samples on the same chromosome.

So the plan is when incorporating variants into the 3-frame TVG, adjacent variants are identified and artificial MNV variants are created. We will also use a user defined cutoff to control how many variants we should consider to merge as MNV and set default to 2, meaning that we won't consider MNVs with 3 or more adjacent SNPs.