[maker-devel] Blastx of repeats with mpi maker failing on small contigs

[Michael Nuhn]

Hello Carson!

(Mpi) Maker (2.27) is failing when it runs blast searches.

It prints out the command it is trying to run. When I try to run this 
command manually on the command line, blast terminates with an error, 
because it either can't find the input file or it can't find a file 
ending in .pin, which I think is a protein index file it expects to be 
there.

I've looked at a few contigs on which maker fails and they were all 
rather short contigs.

Maker works fine, if I

- run it without mpi or
- run it with mpi, but a maximum of 4 processors.

(Mpi) Maker used to run fine with 128 processors before this.

The contigs are sorted descending by size in the genome file. I think 
maker has processed the large ones and the problems it is having now 
might have something to do with it running on smaller contigs.

 From looking at the error messages I thought at first the index file of 
the genome might be corrupted, so I deleted it and let maker rebuild it. 
This didn't fix the issue though. I have also set the path for temporary 
files manually to make sure maker is not running out of temporary space.

Any idea how to overcome this?.

Cheers,
Michael.

P.S.: A typical error message I'm getting is this:

--Next Contig--

[blastall] FATAL ERROR:  search cannot proceed due to errors in all 
contexts/frames of query sequences
running  blast search.
#--------- command -------------#
Widget::blastx:
/nfs/panda/ensemblgenomes/external/blast/bin/blastall -p blastx -d 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/te_proteins%2Efasta.mpi.10.0 

-i 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/rank16/LSalAtl2s8087.0 
-b 10000 -v 10000 -e 1e-06 -z 300 -Y 500000000 -a 1 -U -F T -I T -o /n
fs/production/panda/ensemblgenomes/development/mnuhn/Sea_louse/test/maker_final_assembly_III/LSalAtl2s.maker.output/LSalAtl2s_datastore/A2/0B/LSalAtl2s8087//
theVoid.LSalAtl2s8087/LSalAtl2s8087.0.te_proteins%2Efasta.repeatrunner.temp_dir/te_proteins%2Efasta.mpi.10.0.repeatrunner
#-------------------------------#
ERROR: Chunk failed at level:2, tier_type:0
FAILED CONTIG:LSalAtl2s8083

doing blastx repeats
setting up GFF3 output and fasta chunks
doing blastx repeats
re reading repeat masker report.
/nfs/production/panda/ensemblgenomes/development/mnuhn/Sea_louse/test/maker_final_assembly_III/LSalAtl2s.maker.output/LSalAtl2s_datastore/2C/53/LSalAtl2s8249//theVoid.LSalAtl2s8249/LSalAtl2s8249.0.all.rb.out
[blastall] FATAL ERROR:  search cannot proceed due to errors in all 
contexts/frames of query sequences
[blastall] FATAL ERROR:  search cannot proceed due to errors in all 
contexts/frames of query sequences
running  blast search.
running  blast search.
#--------- command -------------#
Widget::blastx:
/nfs/panda/ensemblgenomes/external/blast/bin/blastall -p blastx -d 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/te_proteins%2Efasta.mpi.10.0 
-i 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/rank26/LSalAtl2s8135.0 
-b 10000 -v 10000 -e 1e-06 -z 300 -Y 500000000 -a 1 -U -F T -I T -o 
/nfs/production/panda/ensemblgenomes/development/mnuhn/Sea_louse/test/maker_final_assembly_III/LSalAtl2s.maker.output/LSalAtl2s_datastore/EF/10/LSalAtl2s8135//theVoid.LSalAtl2s8135/LSalAtl2s8135.0.te_proteins%2Efasta.repeatrunner.temp_dir/te_proteins%2Efasta.mpi.10.0.repeatrunner
#-------------------------------#
#--------- command -------------#
Widget::blastx:
/nfs/panda/ensemblgenomes/external/blast/bin/blastall -p blastx -d 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/te_proteins%2Efasta.mpi.10.0 
-i 
/nfs/nobackup2/ensemblgenomes/mnuhn/maker/temp2/maker_u5Dl1K/rank19/LSalAtl2s8119.0 
-b 10000 -v 10000 -e 1e-06 -z 300 -Y 500000000 -a 1 -U -F T -I T -o 
/nfs/production/panda/ensemblgenomes/development/mnuhn/Sea_louse/test/maker_final_assembly_III/LSalAtl2s.maker.output/LSalAtl2s_datastore/CA/2E/LSalAtl2s8119//theVoid.LSalAtl2s8119/LSalAtl2s8119.0.te_proteins%2Efasta.repeatrunner.temp_dir/te_proteins%2Efasta.mpi.10.0.repeatrunner
#-------------------------------#
[blastall] FATAL ERROR:  search cannot proceed due to errors in all 
contexts/frames of query sequences
#---------------------------------------------------------------------
Now retrying the contig!!
SeqID: LSalAtl2s8449
Length: 2187
Tries: 18!!
#---------------------------------------------------------------------