From code review:
  Get rid of "UCSC" in variable names (was actually dead code anyway).
  Change scoring function.

as we think it has before trying to INSERT them.

Correctly create an 'analysis' entry if one is missing.

Removed a lot of spaces and tabs at the end of lines.

Squeezed multiple blank lines.

    my @edits = ('initial_met', '_selenocysteine', 'amino_acid_sub');
for Drosophila.  (Like Selenocysteine, but any amino acid including '*' can be substituted.)

added -nonreleasemode to the list of options. This means the update does not require the master schema database and if automatically forced

$do_upload is true.

TODO: testing

of the cases when an appropriate analysis_id does not exist (create a
new analysis).

TODO: Upload.

Add "Was not best match" to unmapped reasons.
Sort unmapped reasons.

TODO:

  1. Pick only the best Xrefs of those that pass the thresholds.
  2. Store new analysis (if needed).
  3. Do the upload of the tables.

Slight formatting.

Still to do:

  1. Write 'unmapped_reason'.
  2. Write 'unmapped_object'.
  3. Upload.