IMP: Expand alphabet for ProteinFASTAFormat

[Sann5]

Context

• Currently, the allowed alphabet for files defined as ProteinFASTAFormat is the alphabet in upper case (excluding J,O andU) plus the * character.

Proposed update

1. Include U, the one-letter symbol for Selenocysteine, recognized by IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN).
2. Include O, the one-letter symbol for Pyrrolysine, recommended since 2009 by the IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN), but not officially included in the standard one-letter code symbols.
3. Include J, the one-letter symbol representing Leucine (L) or Isoleucine (I).
   • B and Z are already used to encode ambiguity between amino acids, so it would be natural to include J as well.
   • Even thou the use of J is still not part of the IUPAC standard, it is already commonly found in public databases. For example, this article claims that the RefSeq non-redundant proteins contained as of 2013, 15309 occurrences of the character J.
   • It was also been endorsed by institutions active in the bioinformatics community:
     • MIT
     • DNA Data Bank of Japan
4. Finally, I would like the format to allow for lowercase symbols.
   • Some use lowercase symbols to denote stereoisomerism.
   • Furthermore, most tools are case insensitive, effectively mapping lower-case symbols to their upper-case counterparts.
     • Supporting evidence A: Yang Zhang's Research Group, FASTA format definition, Department of Computational Medicine and Bioinformatics, University of Michigan Medical School
     • Supporting evidence B: some tool by the University of Maryland.
   • Therefore I do not think that this will lead to undefined behavior with most tools, but of course, the only way to be sure is to try or check with each tool (something I would very much like to avoid).

Motivation

I'm trying to use a database of protein sequences that contain all of the above. This is part of some enhancements I'm working on for the q2-moshpit plugin.

[Sann5]

Thank you @gregcaporaso for the detailed review. I will make the proposed changes and re-request a review.

• [x] Implement MixedCaseProteinFASTAFormat
• [x] Expand the test sequences for the ProteinFASTAFormat (to include the new proposed characters)
• [x] Create additional tests for the MixedCaseProteinFASTAFormat

I ended up adding a MixedCaseProteinSequencesDirectoryFormat and a MixedCaseAlignedProteinSequencesDirectoryFormat (and some tests for them). Let me know if I should remove this.

[gregcaporaso]

@Sann5, just FYI we'll be getting back to this for a re-review shortly. Thanks for the updates!

[lizgehret]

Hi @Sann5,

Thanks for your patience! I'll review this today and follow up with any questions or requested changes.

[lizgehret]

On first glance, this all looks quite reasonable! I'm going to pull your changes down locally to test - unless anything unusual comes up, this should be good to go.

[Sann5]

Thank you for the thorough review @lizgehret. Cheers!

[gregcaporaso]

@lizgehret, @Sann5 - I notice that all 26 english letters are now in this alphabet. Is that the intention? I just want to confirm that this doesn't contain any unintended characters. Scratch that, it seems right: 20 standard amino acids, plus X, U, O, J, B, Z = 26.

Thanks for the work on this one!