rsamtools scanBcf segfault

0

Entering edit mode

Rob Syme ▴ 20

@rob-syme-4776

Last seen 3.6 years ago

Canada

Hi all, I'm looking to read in a bcf file, but R keeps segfaulting out. I'd love advice or pointers if anyone has seen this before. Environment: Rsamtools version 1.4.2 R version 2.13.1 (2011-07-08) Code to reproduce is?gist.github.com/1109210 reproduced below: wget 08ae7fe9927/subset.vcf grep -v "^#" subset.vcf | cut -f 1 | sort -u > subset.dict bcftools view -D subset.dict -bS subset.vcf > subset.bcf bcftools index subset.bcf wget 3cdd942c083/segfault.R Rscript segfault.R ## END BASH I get: $ Rscript segfault.R ... *** caught segfault *** address (nil), cause 'memory not mapped' Traceback: 1: .Call(func, .extptr(file), space, tmpl) 2: doTryCatch(return(expr), name, parentenv, handler) 3: tryCatchOne(expr, names, parentenv, handlers[[1L]]) 4: tryCatchList(expr, classes, parentenv, handlers) 5: tryCatch({ .Call(func, .extptr(file), space, tmpl)}, error = function(err) { stop("scanBcf: ", conditionMessage(err), "\n path: ", path(file), call. = FALSE)}) 6: .io_bcf(.scan_bcf, file, ..., param = param) 7: .local(file, ...) 8: scanBcf(bcf, param = param) 9: scanBcf(bcf, param = param) aborting ... Segmentation fault

• 1.4k views

ADD COMMENT • link updated 13.5 years ago by Martin Morgan 25k • written 13.5 years ago by Rob Syme ▴ 20

0

Entering edit mode

Martin Morgan 25k

@martin-morgan-1513

Last seen 15 days ago

United States

On 07/27/2011 05:17 AM, Rob Syme wrote: > Hi all, > I'm looking to read in a bcf file, but R keeps segfaulting out. I'd > love advice or pointers if anyone has seen this before. > Environment: > Rsamtools version 1.4.2 > R version 2.13.1 (2011-07-08) > > Code to reproduce is gist.github.com/1109210 reproduced below: > > wget 6608ae7fe9927/subset.vcf > grep -v "^#" subset.vcf | cut -f 1 | sort -u> subset.dict > bcftools view -D subset.dict -bS subset.vcf> subset.bcf > bcftools index subset.bcf > wget 813cdd942c083/segfault.R > Rscript segfault.R > ## END BASH > I get: Thanks Rob for the nice reproducible example. The problem is when you open the BcfFile; 'mode' should be 'rb' to indicate that you want to read a binary file. Or leave it unspecified and the heuristic will do the right thing. I'll try to get a fix in so that R doesn't seg fault. Martin > > $ Rscript segfault.R > ... > *** caught segfault *** > address (nil), cause 'memory not mapped' > > Traceback: > 1: .Call(func, .extptr(file), space, tmpl) > 2: doTryCatch(return(expr), name, parentenv, handler) > 3: tryCatchOne(expr, names, parentenv, handlers[[1L]]) > 4: tryCatchList(expr, classes, parentenv, handlers) > 5: tryCatch({ .Call(func, .extptr(file), space, tmpl)}, error = > function(err) { stop("scanBcf: ", conditionMessage(err), "\n path: > ", path(file), call. = FALSE)}) > 6: .io_bcf(.scan_bcf, file, ..., param = param) > 7: .local(file, ...) > 8: scanBcf(bcf, param = param) > 9: scanBcf(bcf, param = param) > aborting ... > Segmentation fault > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- Computational Biology Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109 Location: M1-B861 Telephone: 206 667-2793

ADD COMMENT • link 13.5 years ago Martin Morgan 25k

0

Entering edit mode

> Thanks Rob for the nice reproducible example. The problem is when you open > the BcfFile; 'mode' should be 'rb' to indicate that you want to read a > binary file. Or leave it unspecified and the heuristic will do the right > thing. Thanks Martin, the manual at http://www.bioconductor.org/packages/2.8/bioc/manuals/Rsamtools/man/Rs amtools.pdf includes a minor error that threw me off. The "Usage" section is correct (and I should have read this more closely), but in the "Arguments" section, the mode argument is described as: "A character(1) vector; mode="rw" indicates a binary (BCF) ?le, mode="r" a text (VCF) ?le" when it should probably be "A character(1) vector; mode="rb" indicates a binary (BCF) ?le, mode="r" a text (VCF) ?le" I have another error - and I've had a closer read of the manual this time ;) I have generated a filtered vcf using freebayes, but the following commands throw an error: wget http://gist.github.com/raw/1111101/f8629dfa20b15893684c78a586f321 dbc65c6b6c/subset.vcf grep -v "^#" subset.vcf | cut -f 1 | sort -u > subset.dict bcftools view -bSD subset.dict subset.vcf > subset.bcf bcftools index subset.bcf wget http://gist.github.com/raw/1111101/5684d61541b39fb8c02e765d5f7891 d9d122244e/scanBcf_error.R Rscript scanBcf_error.R # This gives the error: Error: scanBcf: failed to find fmt encoded as '21057' path: <<redacted>>/subset.bcf Execution halted # Calling scanBcfHeader also gives an error: scanBcfHeader(bcf) Error in mapply(f, ..., SIMPLIFY = FALSE) : 'names' attribute [5] must be the same length as the vector [4] I can't see anything wrong with the VCF, am I missing something obvious again?

ADD REPLY • link 13.5 years ago Rob Syme ▴ 20

0

Entering edit mode

On 07/28/2011 12:08 AM, Rob Syme wrote: >> Thanks Rob for the nice reproducible example. The problem is when you open >> the BcfFile; 'mode' should be 'rb' to indicate that you want to read a >> binary file. Or leave it unspecified and the heuristic will do the right >> thing. Thanks that'll be corrected. > > Thanks Martin, the manual at > http://www.bioconductor.org/packages/2.8/bioc/manuals/Rsamtools/man/ Rsamtools.pdf > includes a minor error that threw me off. The "Usage" section is > correct (and I should have read this more closely), but in the > "Arguments" section, the mode argument is described as: > "A character(1) vector; mode="rw" indicates a binary (BCF) ?le, > mode="r" a text (VCF) ?le" > when it should probably be > "A character(1) vector; mode="rb" indicates a binary (BCF) ?le, > mode="r" a text (VCF) ?le" > > > I have another error - and I've had a closer read of the manual this time ;) > I have generated a filtered vcf using freebayes, but the following > commands throw an error: > > wget http://gist.github.com/raw/1111101/f8629dfa20b15893684c78a586f3 21dbc65c6b6c/subset.vcf > grep -v "^#" subset.vcf | cut -f 1 | sort -u> subset.dict > bcftools view -bSD subset.dict subset.vcf> subset.bcf > bcftools index subset.bcf > wget http://gist.github.com/raw/1111101/5684d61541b39fb8c02e765d5f78 91d9d122244e/scanBcf_error.R > Rscript scanBcf_error.R > > # This gives the error: > Error: scanBcf: failed to find fmt encoded as '21057' > path:<<redacted>>/subset.bcf > Execution halted The problem is that the current implementation isn't really general, it implements the formats supported (computed on, rather than just echoed) by bcftools -- PL, DP, GQ, SP, GT, GL -- whereas vcf is really much more flexible than that. This is cryptically documented on ?scanBcf, but I'd like to make this more flexible and will work on this. > # Calling scanBcfHeader also gives an error: > scanBcfHeader(bcf) > Error in mapply(f, ..., SIMPLIFY = FALSE) : > 'names' attribute [5] must be the same length as the vector [4] This has been addressed previously in the devel branch. Thanks Rob for the reports. Martin > > > I can't see anything wrong with the VCF, am I missing something obvious again? -- Computational Biology Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109 Location: M1-B861 Telephone: 206 667-2793

ADD REPLY • link 13.5 years ago Martin Morgan 25k

Login before adding your answer.