improvements for large genome downloads

[bluegenes]

I'm getting some unexpected failures trying to download some large plant genomes. My guess is that these downloads take a lot longer, so I need to modify parameters to avoid blocking by NCBI.

To fix:

• could make # semaphore permits a modifiable parameter (currently 3, but maybe use 1 for large genomes).
• can we use an API key to get a higher download limit?
• need to double check that we're not attempting protein downloads when we're not keeping fastas + not generating protein signatures. Despite tests and outputs, I do see some "protein" filenames in the failures file, so let's check again. Attempting these downloads might be the cause of hitting ncbi download limits.