[maker-devel] Odd missing label bug

[Fields, Christopher J]

I’m seeing an odd bug in GFF output where the label is being sporadically left off some child ‘match_part' features.  This is when using maker-2.30p.  I found this while splitting out the data by source as individual tracks for IGV viewing; I have several sets of ESTs from individuals of the same species, so I have them labeled (from maker_opts):

est=./ESTs/Trinity_1128_v2_combined_pr2.clean.fasta:strain_1128
protein=./ESTs/xeno/Taeniopygia_guttata.taeGut3.2.4.73.pep.all.fa:TaeGut,./ESTs/xeno/Ficedula_albicollis.FicAlb_1.4.73.pep.all.fa:FicAlb

I planned on separating these into separate tracks, one for each BLASTX, BLASTN, etc.  

After a test run on a few scaffolds, a small number of the BLASTX and BLASTN look like this:

KB913263.1      blastx:TaeGut   protein_match   374736  401929  330     -       .       ID=KB913263.1:hit:9:3.10.0.3;Name=ENSTGUP00000000074
KB913263.1      blastx  match_part      401756  401929  312     -       .       ID=KB913263.1:hsp:55:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M58;Target=ENSTGUP00000000074 1 58
KB913263.1      blastx  match_part      394486  394665  186     -       .       ID=KB913263.1:hsp:56:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M42 D9 M9;Target=ENSTGUP00000000074 56 106
KB913263.1      blastx  match_part      392059  392178  204     -       .       ID=KB913263.1:hsp:57:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M40;Target=ENSTGUP00000000074 97 136
KB913263.1      blastx  match_part      387479  387547  116     -       .       ID=KB913263.1:hsp:58:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M23;Target=ENSTGUP00000000074 151 173
KB913263.1      blastx  match_part      383644  383742  156     -       .       ID=KB913263.1:hsp:59:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M33;Target=ENSTGUP00000000074 172 204
KB913263.1      blastx  match_part      382860  383063  242     -       .       ID=KB913263.1:hsp:60:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M68;Target=ENSTGUP00000000074 193 260
KB913263.1      blastx  match_part      382538  382648  186     -       .       ID=KB913263.1:hsp:61:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M37;Target=ENSTGUP00000000074 247 283
KB913263.1      blastx  match_part      382072  382236  268     -       .       ID=KB913263.1:hsp:62:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M55;Target=ENSTGUP00000000074 282 336
KB913263.1      blastx  match_part      379270  379458  330     -       .       ID=KB913263.1:hsp:63:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M63;Target=ENSTGUP00000000074 336 398
KB913263.1      blastx  match_part      374736  374900  147     -       .       ID=KB913263.1:hsp:64:3.10.0.3;Parent=KB913263.1:hit:9:3.10.0.3;Gap=M55;Target=ENSTGUP00000000074 4 58

Note the missing label in the source column for the ‘match_part’ types.  Strangely, only a small # have this problem, but the exact same ones appear to be consistently popping up on repeated runs (this is from a test prior to a full launch using openmpi).  It does seem like only child ‘match_part’ appears affected.

Any ideas?

chris