Parameterize the database used for checking if sequences are from fragment transcripts that encode long protein coding genes

[taylorreiter]

So that the demo data will run faster. We could supply this as a URL in the config, along with the database name. Or, we could supply it as a URL in the config and then make references to the database generic. It's a 12 gb fasta so downloading it takes a bit, and then searching it also takes a few minutes, so running on the demo data is a bit slow. Tolerable, but would probably be better over the long term to parameterize. Not prioritizing this bc I think uniref50 is a great default for most peptigate runs.