[Bioc-devel] ShortRead readFasta UniProt Incorrect Import

[Dario Strbenac]

Good day,

If I have a FASTA file that contains

>sp|Q9NYW0|T2R10_HUMAN Taste receptor type 2 member 10 OS=Homo sapiens GN=TAS2R10 PE=1 SV=3
MLRVVEGIFIFVVVSESVFGVLGNGFIGLVNCIDCAKNKLSTIGFILTGLAISRIFLIWI
IITDGFIQIFSPNIYASGNLIEYISYFWVIGNQSSMWFATSLSIFYFLKIANFSNYIFLW
LKSRTNMVLPFMIVFLLISSLLNFAYIAKILNDYKTKNDTVWDLNMYKSEYFIKQILLNL
GVIFFFTLSLITCIFLIISLWRHNRQMQSNVTGLRDSNTEAHVKAMKVLISFIILFILYF
IGMAIEISCFTVRENKLLLMFGMTTTAIYPWGHSFILILGNSKLKQASLRVLQQLKCCEK
RKNLRVT

readFasta fails to import it with the warning

proteins <- readFasta('.', "test.fasta")

Warning message:
In .Call2("fasta_index", filexp_list, nrec, skip, seek.first.rec,  :
  reading FASTA file test.fasta: ignored 129 invalid one-letter sequence codes

Also, the amino acid sequence is incomplete. There are 308 amino acids, but 

> width(proteins)
[1] 178

It's undesirable for users that some amino acids are discarded. Hopefully, they notice the warning message before proceeding with the analysis.

Admittedly, readFasta is in ShortRead, so is designed to work with high througput sequencing reads. But, perhaps it would be better suited to a infrastructure package such as Biobase and generalised to correctly import any FASTA file. There's even a Bioconductor workflow at https://www.bioconductor.org/help/workflows/sequencing/ which has a section titled "DNA/amino acid sequence from FASTA files" and demonstrates the use of readFasta.

I used version 1.34.2 of ShortRead which is the newest one.

--------------------------------------
Dario Strbenac
University of Sydney
Camperdown NSW 2050
Australia