Assertion failure when mapping barley genome

[glennhickey]

This is courtesy of @LindaMilne via ComparativeGenomicsToolkit/cactus#800 and reproduces with the minigraph-0.19_x64-linux release binary:

wget https://ics.hutton.ac.uk/barleyrtd/data/barley_two_old_genomes.gfa.gz
wget https://ics.hutton.ac.uk/barleyrtd/data/morex_1H_old.fasta.gz
minigraph barley_two_old_genomes.gfa.gz morex_1H_old.fasta.gz -xasm -c -t 32 -o morex.gaf

minigraph [M::main::7.941*1.00] loaded the graph from "barley_two_old_genomes.gfa.gz"                                                                                                                                                                                                                                                       [M::mg_index::59.948*1.27] indexed the graph
[M::mg_opt_update::62.273*1.26] occ_max1=100; lc_max_occ=2
minigraph: galign.c:133: mg_gchain_cigar: Assertion `l == gc->qe - gc->qs && gc->p->aplen == gc->pe - gc->ps' failed.
Aborted

[glennhickey]

@lh3 this issue seems to have arisen again in cow genomes: ComparativeGenomicsToolkit/cactus#832 (comment)

RuntimeError: Command /usr/bin/time -v bash -c 'set -eo pipefail && minigraph /tmp/664b8033018b57ed8eeaf89049092aea/35c4/4b7d/tmpdsu3dzqo/mg.gfa /tmp/664b8033018b57ed8eeaf89049092aea/35c4/4b7d/tmpdsu3dzqo/Duroc.fa -o /tmp/664b8033018b57ed8eeaf89049092aea/35c4/4b7d/tmpdsu3dzqo/Duroc.gaf -c -xasm -t 30' exited 134: stdout=None, stderr=[M::main::0.4641.00] loaded the graph from "/tmp/664b8033018b57ed8eeaf89049092aea/35c4/4b7d/tmpdsu3dzqo/mg.gfa"
[M::mg_index::7.9171.58] indexed the graph
[M::mg_opt_update::8.4481.55] occ_max1=100; lc_max_occ=2
minigraph: galign.c:133: mg_gchain_cigar: Assertion `l == gc->qe - gc->qs && gc->p->aplen == gc->pe - gc->ps' failed.
Command terminated by signal 6

[lh3]

I can see a download link in ComparativeGenomicsToolkit/cactus#800. What's the command line in use?

[glennhickey]

For the barley genome, the command line is above:

wget https://ics.hutton.ac.uk/barleyrtd/data/barley_two_old_genomes.gfa.gz
wget https://ics.hutton.ac.uk/barleyrtd/data/morex_1H_old.fasta.gz
minigraph barley_two_old_genomes.gfa.gz morex_1H_old.fasta.gz -xasm -c -t 32 -o morex.gaf

If I remember, it takes about an hour and (maybe around 40Gb RAM) to crash. The gfa would have been constructed with minigraph -cxggs for those genomes.

Thanks!

[lh3]

PS: just cut release v0.20 for this and a few other bug fixes. Results should remain the same.

[glennhickey]

Awesome, thanks!!