glennhickey
commented
1 year ago That is very strange. My only guess is that hg38.nrs.paf is somehow corrupt. Can you share the output of tail hg38.nrs.paf ?
Open dudududu12138 opened 1 year ago
glennhickey
commented
1 year ago That is very strange. My only guess is that hg38.nrs.paf is somehow corrupt. Can you share the output of tail hg38.nrs.paf ?
dudududu12138
commented
1 year ago That is very strange. My only guess is that
hg38.nrs.pafis somehow corrupt. Can you share the output oftail hg38.nrs.paf?
Thank you for your reply.
This is my output of tail hg38.nrs.paf
id=Ref|HLA-B*67:01:02 2675 0 2675 + id=_MINIGRAPH_|s3148 2675 0 2675 2675 2675 60 cg:Z:2675M
id=Ref|HLA-B*82:02:01 3050 0 3050 + id=_MINIGRAPH_|s3153 3050 0 3050 3050 3050 60 cg:Z:3050M
id=Ref|HLA-C*01:21 2895 0 2895 + id=_MINIGRAPH_|s3162 2895 0 2895 2895 2895 60 cg:Z:2895M
id=Ref|HLA-C*02:10 2893 0 2893 + id=_MINIGRAPH_|s3167 2893 0 2893 2893 2893 60 cg:Z:2893M
id=Ref|HLA-C*08:04:01 2895 0 2895 + id=_MINIGRAPH_|s3256 2895 0 2895 2895 2895 60 cg:Z:2895M
id=Ref|HLA-C*12:13 3058 0 3058 + id=_MINIGRAPH_|s3271 3058 0 3058 3058 3058 60 cg:Z:3058M
id=Ref|HLA-C*12:22 2895 0 2895 + id=_MINIGRAPH_|s3273 2895 0 2895 2895 2895 60 cg:Z:2895M
id=Ref|HLA-C*15:13 2895 0 2895 + id=_MINIGRAPH_|s3282 2895 0 2895 2895 2895 60 cg:Z:2895M
id=Ref|HLA-C*15:16 3066 0 3066 + id=_MINIGRAPH_|s3283 3066 0 3066 3066 3066 60 cg:Z:3066M
id=Ref|HLA-C*16:02:01 2895 0 2895 + id=_MINIGRAPH_|s3287 2895 0 2895 2895 2895 60 cg:Z:2895MThis is the top5 headers of my output in file hg38.nrs.sv.gfa.fa.gz.
>id=_MINIGRAPH_|s1
>id=_MINIGRAPH_|s2
>id=_MINIGRAPH_|s3
>id=_MINIGRAPH_|s4
>id=_MINIGRAPH_|s5This is my seqfile.
(Ref:1.0,seq1:1.0,_MINIGRAPH_:1.0);
Ref MyPath/hg38.fa
seq1 MyPath/seq1.fa
_MINIGRAPH_ file:///MyPath/hg38.nrs.sv.gfa.fa.gzhmm, I wonder if the *'s in the names are causing problems... Are you able to share the input to cactus-align so I can try to reproduce?
dudududu12138
commented
1 year ago hmm, I wonder if the *'s in the names are causing problems... Are you able to share the input to
cactus-alignso I can try to reproduce?
Oh,thanks!!! But the hg38.nrs.sv.gfa.fa.gz file is very large and I can't upload to the public website. So I share the Reference sequence and my seq1.You can reproduce from the first step. This is my reference fasta file: Reference This is my seq1 fasta file: seq1.fa This is my seqfile: seqfile.txt This is the paf file produeced in step 2: paf
glennhickey
commented
1 year ago I will take a look. But please be aware that Cactus is not designed to work on this type of data. See here for an explanation: https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#hprc-graph-setup-and-name-munging
As mentioned in that link, if you want to have multiple copies of the same allele, they need to be associated with different samples. This should probably be more prominently warned in the documentation. I would like to make an example of how to do this with GRCh38 but have not yet had enough time / volunteers. But for now, if you want to make an HLA pangenome, you need a separate fasta (and seqfile line) for each allele.
dudududu12138
commented
1 year ago I will take a look. But please be aware that Cactus is not designed to work on this type of data. See here for an explanation: https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#hprc-graph-setup-and-name-munging
As mentioned in that link, if you want to have multiple copies of the same allele, they need to be associated with different samples. This should probably be more prominently warned in the documentation. I would like to make an example of how to do this with GRCh38 but have not yet had enough time / volunteers. But for now, if you want to make an HLA pangenome, you need a separate fasta (and seqfile line) for each allele.
I will take a look. But please be aware that Cactus is not designed to work on this type of data. See here for an explanation: https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#hprc-graph-setup-and-name-munging
As mentioned in that link, if you want to have multiple copies of the same allele, they need to be associated with different samples. This should probably be more prominently warned in the documentation. I would like to make an example of how to do this with GRCh38 but have not yet had enough time / volunteers. But for now, if you want to make an HLA pangenome, you need a separate fasta (and seqfile line) for each allele.
Hello, sorry for the late reply. I dropped the HLA sequences and constructed the pangenome. The previously reported error is resolved. But a new error has appeared. This error also occurs at the cactus-align step.
My error log file is listed here:
logError.txt.
To solve this new error, I renamed my fasta files. But it didn't work. The top5 headers of my reference fasta file and my NRS fasta file are listed below:
>chr1
>chr2
>chr3
>chr4
>chr5>NRS#1#WGC012904D_622982
>NRS#1#WGC012904D_5534358
>NRS#1#WGC012904D_9614265
>NRS#1#WGC012904D_1067906
>NRS#1#WGC012904D_5090934Thank you for your help!
glennhickey
commented
1 year ago This error, segfault in hal2vg, seems like a bug. I'd imagine that splitting by chromosome (see the documentation) would resolve this. Still, I'd be curious to know why it's crashing. Are you able to share "out.hal" with me somehow so I can reproduce?
(you would need to rerun your failing command with --disableCaching --cleanWorkDir never --restart to save out.hal)
dudududu12138
commented
1 year ago This error, segfault in hal2vg, seems like a bug. I'd imagine that splitting by chromosome (see the documentation) would resolve this. Still, I'd be curious to know why it's crashing. Are you able to share "out.hal" with me somehow so I can reproduce?
(you would need to rerun your failing command with
--disableCaching --cleanWorkDir never --restartto saveout.hal)
Hello,I tried with your suggesstion to get the .hal file. But it report an error. This is my code:
singularity exec $sif cactus-align \
--disableCaching --cleanWorkDir never --restart \
./jobstore seqfile.txt hg38.nrs.paf hg38.nrs.hal \
--workDir $TEMP --logFile tmp.log \
--pangenome --outVG --reference Ref \
--defaultCores 40 --defaultMemory 160G --defaultDisk 200GAnd the error log file is here: newlogError.txt.
By the way,what do you mean "splitting by chromosome"? I only have two fasta files, one for GRCh38 and one for NRS sequence, and my goal is to insert NRS (non reference sequence) sequence into GRCh38 to construct the graph pan-genome. If I split GRCh38 by chromosome, should I align my NRS sequence with all the chromosomes?
glennhickey
commented
1 year ago The only way that ocmmand will work is if you mount jobstore and workdir with singularity using -b.
For chromosome splitting, see the example https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#yeast-splitting-by-chromosome and https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#hprc-graph-splitting-by-chromosome
dudududu12138
commented
1 year ago The only way that ocmmand will work is if you mount jobstore and workdir with singularity using -b.
For chromosome splitting, see the example https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#yeast-splitting-by-chromosome and https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md#hprc-graph-splitting-by-chromosome
Sorry, I am confused with your advice. What's the meaning of "using -b"? Do you mean the parameter --binariesMode in cactus-align?
glennhickey
commented
1 year ago Sorry, I meant -B. Something like singularity exec -B /home/dudu/cactus:/data cactus-align /data/jobstore --workDir /data --disableCacing --cleanWorkDir never ....
dudududu12138
commented
1 year ago I
Sorry, I meant
-B. Something likesingularity exec -B /home/dudu/cactus:/data cactus-align /data/jobstore --workDir /data --disableCacing --cleanWorkDir never ....
Hello, I followed your advice. But I still got error message. This is my codes:
singularity exec -B MyPath:/data $sif cactus-align \
--disableCaching --cleanWorkDir never --restart \
/data/jobstore seqfile.txt hg38.nrs.paf hg38.nrs.hal \
--workDir /data --logFile tmp3.log \
--pangenome --outVG --reference Ref \
--defaultCores 40 --defaultMemory 160G --defaultDisk 200GAnd this is the error message: logFile.txt.
jiadong324
commented
9 months ago I run into similar errors for command gaffilter.
Here is a part of my seqfile.txt.
CHM13_h1 chr22_42590000-42655000.fa
HG00171_h1 HG00171_h1-haplotype1-0000002.fa
HG00171_h2 HG00171_h2-haplotype2-0000152.fa
HG00268_h1 HG00268_h1-haplotype1-0000005.fa
HG00268_h2 HG00268_h2-haplotype2-0000050.faThis is the error message.
RuntimeError: Command ['singularity', '--silent', 'exec', '-u', '-B', '/tmp/d6343a4bdf455ba6bda4cfd0acfa3261/d377/8eb2/tmpdcj3uq7z:/mnt', '--pwd', '/mnt', '/hgsvc_chr22_assemblies/mc/jobStore/cactus.img', 'gaffilter', 'mg.paf.filter', '-p', '-r', '5.0', '-m', '0.0', '-b', '250000', '-q', '5', '-i', '0.5'] signaled SIGABRT: stderr=gaffilter: paf.hpp:75: PafLine parse_paf_line(const string&): Assertion `tag_toks.size() == 3' failed.Here is the PAF file, and it looks good. mc_hgsvc_cyp2d6.paf.zip
Hello, I constructed the graph pangenome according to the process (https://github.com/ComparativeGenomicsToolkit/cactus/blob/master/doc/pangenome.md (the quick start section)) you gave. The first two steps are ok.(1.cactus-minigraph 2.cactus-graphmap) But the third step report an error. My cactus version is v2.4.2 My codes are listed below:
The error messages are listed below: