Handle MOIRE "NA values in allele frequency vector" warnings

[a-hubbard]

When run on example2_allele_table.tsv, the MOIRE wrapper generates a number of warnings that take the form:

NA values detected in allele frequency vector. This may indicate a problem with the MCMC chain or there are loci with no diversity. NA values will be replaced with 0.

This probably does not indicate an issue, as it is likely that the loci in question do in fact lack diversity (as Jorge and Alfred H. discussed at the hackathon). This is not urgent, but we should add code to 1) capture these warnings, 2) check if they can all be explained by a lack of diversity, and 3) don't pass the warnings onward if so.

Alternatively, we could ask Max to update MOIRE so that it checks for a lack of diversity in these cases within the package, and only warns if that can't explain the problem.

[jorgeamaya]

Stand-by. Overlap with issue #45

[jorgeamaya]

Note:

In some cases, Moire infers uncommon alleles are false positives. Loci which become monoallelic after removing uncommon alleles have he "0" in he_summary. Truly monoallelic loci are not printed by default by moire, but we'e modifying the output table to include them, manually assigning 0 values.