Consistency between NCBI taxonomy versions in the kraken and the custom bowtie databases

[mikemc]

Since NCBI taxonomy evolves over time (e.g.), it's possible that the NCBI taxonomy at the time that the index workflow is run will differ from that that was used in whatever kraken database is being pulled in the index workflow.

How I'd imagine handling this would be to do something like: make sure we're using a GenBank release that is compatible with whatever RefSeq release is used in the Kraken database, and then make sure that we have the correct NCBI taxonomy dump for these releases.

Alternatively, there are probably ways to determine what the changes are, but it might not be easy to figure out what to do with the differences.

@willbradshaw , can you say something as to how discrepancies are/aren't handled currently?

[willbradshaw]

I think this is essentially a documentation question, rather than a pipeline feature per se. It's up to the user generating the index to make sure that the versions of the various resources they're pulling in are compatible. I agree that the documentation should reflect this, though.

[jeffkaufman]

It's up to the user generating the index

Is this something you and/or Harmon are currently doing when you create indexes?

[willbradshaw]

Not as diligently as I'd like. We'd probably do it better if a procedure was laid out properly in the documentation as suggested.