What is the reason for only including snps in HoneyBADGER? about honeybadger

[JEFworks]

It's a great question. Like snps, indels may also be informative of deletions and LOHs by the same logic. However, we did not have a good way to account for the different lengths of indels, both in the Bayesian hierarchical framework and also in terms of modeling how they may be impacted by mono-allelic detection. If you can show that length does impact detection sensitivity, then presumably indels could be directly incorporated into the same model as what is currently used for snps. Alternatively, if there are sufficient numbers of indels of each differing length, then the rate of mono-allelic detection could be fitted to be a function of of length.

[JEFworks]

We would be happy to accept pull requests and contributions if you are interested in tackling this question.