False positive recombinants

[jeromekelleher]

There seems to be a few cases in the current large ARG where we inferring recombinants due to clusters of adjacent mutations recurring, e.g.,

Note 22030--22034 all mutating together here. I'm not sure what the underlying reason for this is, but there's definitely a bunch of recombinants that have a few hundred descendants that are of these form. One example has a whole bunch of recombinant children, which is definitely suspicious:

Something to investigate for the paper. We should presumably exclude these from analyses as false positives.