UniProtKB gpad to WormBase .ace

• The gpad file that contains all of the C. elegans annotations currently in Protein2GO is produced by UniProtKB on a weekly basis.
  • A new file is available every Monday.
  • The file is located here: ftp://ftp.ebi.ac.uk/pub/contrib/goa/ and is named: gp_association.6239_wormbase.gz

• Download the file from the UniProtKB ftp link and put it on tazendra here (in the appropriate year and month directory):
  • /home/acedb/kimberly/citace_upload/go/gpad2ace
  • for example: /home/acedb/kimberly/citace_upload/go/gpad2ace/2015_February

• To convert the gpad file to a .ace file you'll need:
  • gp2protein.wb file that maps UniProtKB IDs to WBGenes
  • go_gpad_parser.pl

• The go_gpad_parse.pl generates three files:
  • gpad_extra_column - a file that adds the WBGene ID as an extra column (a new column 2) to the gpad file
  • gpad_extra_column.err - a file that indicates:
    • which UniProtKB IDs don't map to WBGene IDs in the gp2protein.wb file, for example:
      • IDs that correspond to a transposon or retrotransposon reverse transcriptase that is encoded in the C. elegans genome (column 3)
        • Q9TZX4
      • IDs that correspond to protein fragments based on translation of incomplete nucleotide submissions to EMBL/GenBank/DDBJ databases (column 3)
        • O17540 - doesn't match 100% to the C. elegans proteome; closest match is kin-20
        • O17541 - near perfect match to kin-24
        • O17542 - closest match is kin-26
        • O17543 - closest match is kin-26
        • O17544 - near perfect match to sid-3
        • O17545 - perfect alignment to kin-32
        • O17546 - perfect alignment to src-1
        • O17911 - perfect alignment to ceh-25 = unc-62
        • O44076 - perfect alignment to cbp-1
        • O44385 - perfect alignment to tat-1 (repeat masked some spans)
        • O61265 - closest match is H19M22.3
      • IDs that correspond to duplicate entries for some full length protein isoforms - these need to be merged in UniProt (column 3)
        • O45049 - lfe-2
        • O45050 - lfe-2
        • O45051 - lfe-2
      • IDs that correspond to full length protein isoforms that only differ between WB and UniProt by one or a few amino acids, i.e. very near match (column 3)
        • O17547 - avr-15
        • O17548 - avr-15
        • O45630 - ncx-2
        • O61304 - abts-1
      • IDs that correspond to non-elegans species, used for ISS annotations or the occasional annotation extension (column 8 or 12)
        • Q8WXF0
    • which PMIDs don't map to a WBPaper ID
    • which annotation extensions can't be mapped to the model
      • Need to add SO_term_relation Text ?SO_term to ?GO_annotation model
    • which UniProtKB IDs are not converted properly in the annotation extensions (column 12)
      • In the script, this corresponds to the code starting at ~line 100
        • This works:
  • gp_annotation.ace - the .ace file for upload to citace

• The gpad file format is documented on the GOC wiki here:
  • http://wiki.geneontology.org/index.php/Final_GPAD_and_GPI_file_format
  • Note that once the parsing script runs and we add an additional column, the column numbers shift one higher in the .err output file

• Possible modifications to pipeline
  • Ask Tony if we could only include in the gpad file annotations to UniProt entries that map to a C. elegans gene ID