Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Segmentation fault while running --meta cov #151

Open
samreenzafer opened this issue Feb 16, 2024 · 0 comments
Open

Segmentation fault while running --meta cov #151

samreenzafer opened this issue Feb 16, 2024 · 0 comments

Comments

@samreenzafer
Copy link

samreenzafer commented Feb 16, 2024

Hi
I have run rvtest several times for single variant and grouped tests, and am now trying to use the --meta score,cov for 3 populations of my cohort separately (Afr, hisp and EURs) - so that I can meta analyze using RAREMETAL.

All my VCFs are divided by Chromosomes and in the correct format expected by rvtest (and I have successfully run the gene based rvtests already on them - example (--inVcf "vcfs/cohort.forRVtest.19.reID.vcf.gz" --out "HISP/lofHC_REVELdamaging_clinvar/maf.0.01//rv.HISP.maf.0.01.kernel.skato.19" --pheno "pheno.HISP.txt" --peopleIncludeFile "HISP/keep.samples.HISP" --peopleExcludeFile "vcfs/exclude_samples.txt" --siteFile "lofHC_REVELdamaging_clinvar.txt.snps" --kernel "skato" --geneFile "refGene_hg38.txt" --freqUpper 0.01 --noweb - which analyzed [INFO] Analyzed [ 2009 ] variants from [ 26969] genes/regions ]

In the similar manner I am now running --meta score,cov subsetting the samples and variants lists in the same fashion, but I get a segmentation fault, and I think this is only happening due the covariance matrix building, Not the Score file. I then ran --meta score and --meta cov separately and the --meta score runs successfully to completion, but the --meta cov does NOT and gives a segmentation fault.

Could this be due to a small number of variants remaining after the variant exclusion criteria (namely, --freqUpper 0.01 --siteFile "lofHC_REVELdamaging_clinvar.txt.snps"), thereby causing a programmatical issue, like NAs or infinite in the matrix? Should I instead try to run the --meta score,cov only on variants filtered by --freqUpper and then when I meta analyze using RAREMETAL, I could try to subset the list of variants to use only the damaging variants I'm interested in ?

Also, there were a handful of chromosomes that did run successfully.
for HISP - chr 11, 14,16,22, & 8 with each analyzing 2146, 954, 1650 ,771 and 1227 variants.
for AFR - chr 13, 15 & 22 with each analyzing 632, 1241, and 745 variants.
for EUR - all chrs ran successfully, with Chr 21 having least # variants analysed = 125, and Chr1 having the largest = 1568.
Looking at this I'm not sure if it's the small number of variants that could be causing a problem.

Here is my successful --meta score command , which shows 324 variants were analyzed.
# ParameterList created by zafers02 on li03c03.chimera.hpc.mssm.edu at Fri Feb 16 12:30:44 2024 --inVcf "/vcfs/cohort.forRVtest.21.reID.vcf.gz" --out "rvtest.HISP.nogene.CategB.21" --pheno "pheno.HISP.txt" --peopleIncludeFile "HISP/keep.samples.HISP" --peopleExcludeFile "vcfs/exclude_samples.txt" --siteFile "lofHC_REVELdamaging_clinvar.txt.snps" --meta "score" --freqUpper 0.01 --noweb [INFO] Parameters END [INFO] Analysis started at: Fri Feb 16 12:30:44 2024 [INFO] Restrict analysis based on specified site file [lofHC_REVELdamaging_clinvar.txt.snps ] [INFO] Loaded [ 184 ] samples from genotype files [INFO] Loaded [ 185 ] sample pheontypes [INFO] Discard [ 1 ] samples as they do not have genotypes [INFO] Loaded 184 male, 0 female and 0 sex-unknonw samples from pheno.HISP.txt [INFO] Loaded 15 cases, 169 controls, and 0 missing phenotypes [WARN] -- Enabling binary phenotype mode -- [INFO] Analysis begins with [ 184 ] samples... [INFO] Impute missing genotype to mean (by default) [INFO] Set upper minor allele frequency limit to 0.01 [INFO] Analysis started **[INFO] Analyzed [ 324 ] variants** [INFO] Analysis ends at: Fri Feb 16 12:33:16 2024

And the header of the Output MetaScore.Assoc.gz is
##ProgramName=Rvtests
##Version=20171009
##Samples=184
##AnalyzedSamples=184
##Families=184
##AnalyzedFamilies=184
##Founders=184
##AnalyzedFounders=184
##InverseNormal=OFF
##TraitSummary min 25th median 75th max mean variance
##Trait 1 1 1 1 2 1.08152 0.0752851
##AnalyzedTrait 0 0 0 0 1 0.0815217 0.0752851
##NullModelEstimates
##Name Beta SD
##Intercept -2.42185 0.0725838
##Sigma2 NA NA

And here is my command --meta cov showing the segmentation fault.
`
Effective Options
--inVcf vcfs/cohort.forRVtest.21.reID.vcf.gz
--out rvtest.HISP.nogene.CategB.cov.21
--pheno pheno.HISP.txt
--peopleIncludeFile HISP/keep.samples.HISP
--peopleExcludeFile vcfs/exclude_samples.txt
--siteFile lofHC_REVELdamaging_clinvar.txt.snps
--meta cov
--freqUpper 0.01
--noweb

[INFO] Program version: 20171009
[INFO] Analysis started at: Fri Feb 16 12:33:50 2024
Include sample [ BKP000684 ].
Include sample [ BKR003225 ].
. . and so on
[INFO] Restrict analysis based on specified site file [ lofHC_REVELdamaging_clinvar.txt.snps ]
[INFO] Loaded [ 184 ] samples from genotype files
[INFO] Loaded [ 185 ] sample pheontypes
[INFO] Discard [ 1 ] samples as they do not have genotypes
[INFO] Loaded 184 male, 0 female and 0 sex-unknonw samples from pheno.HISP.txt
[INFO] Loaded 15 cases, 169 controls, and 0 missing phenotypes
[WARN] -- Enabling binary phenotype mode --
[INFO] Analysis begins with [ 184 ] samples...
[INFO] Meta analysis uses window size 1,000,000 to produce covariance statistics under additive model
[INFO] Impute missing genotype to mean (by default)
[INFO] Set upper minor allele frequency limit to 0.01
[INFO] Analysis started

Segmentation fault
`

I would appreciate any insights into this.
Thanks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant