Bug Bulletin: we have identified a bug that affects indexing when producing gzipped VCFs. This will be fixed in the upcoming 3.2 release; in the meantime you need to reindex gzipped VCFs using Tabix.

arithmetic exception

DavidBJaffeDavidBJaffe Posts: 2Member
edited October 2012 in Ask the team

Dear GATK team,

I'm trying to call variants on some 'haploid' human data (Illumina reads from a mix of clones). I did exactly the following:

crd8% cd /wga/dev/jaffe
crd8% mkdir GATK2
crd8% cd GATK2
crd8% (got GenomeAnalysisTK-2.1-12.tar.bz2 from the GATK site on the web)
crd8% bunzip2 GenomeAnalysisTK-2.1-12.tar.bz2
crd8% cat GenomeAnalysisTK-2.1-12.tar | tar xf -

crd8% cd /local/scratch/jaffe/BroadCRD/fixed

crd8% java -jar /wga/dev/jaffe/GATK2/GenomeAnalysisTK-2.1-12-ga99c19d/GenomeAnalysisTK.jar -R /wga/scr2/bigrefs/human19/genome.fasta -T UnifiedGenotyper -I frag.list -o raw2.vcf -U -baq CALCULATE_AS_NECESSARY -nt 48

crd8% java -jar /wga/dev/jaffe/GATK2/GenomeAnalysisTK-2.1-12-ga99c19d/GenomeAnalysisTK.jar -R /wga/scr2/bigrefs/human19/genome.fasta -T VariantFiltration -U -V raw2.vcf -o var.vcf --filterExpression "QD<5.0||AC<2||DP<6" --filterName junk

The last command failed, with output

... 
INFO  10:38:07,429 TraversalEngine -        Location processed.sites  runtime per.1M.sites completed total.runtime remaining 
WARN  10:38:08,453 Interpreter - ![12,18]: 'QD < 5.0 || AC < 2 || DP < 6;' < error 
java.lang.ArithmeticException: Long coercion: java.util.ArrayList:([1, 1])
    at org.apache.commons.jexl2.JexlArithmetic.toLong(JexlArithmetic.java:914)
    at org.apache.commons.jexl2.JexlArithmetic.compare(JexlArithmetic.java:718)
    at org.apache.commons.jexl2.JexlArithmetic.lessThan(JexlArithmetic.java:774)
    at org.apache.commons.jexl2.Interpreter.visit(Interpreter.java:967)
    at org.apache.commons.jexl2.parser.ASTLTNode.jjtAccept(ASTLTNode.java:18)
...

Can you please suggest a solution? You should be able to access my data, if that would help. Well actually you would have to login to crd8, but you could copy files from there.

Thank you very much.

David

Post edited by Geraldine_VdAuwera on

Best Answer

  • pdexheimerpdexheimer Posts: 299 ✭✭✭
    Answer ✓

    AC contains values for each alternate allele, so is represented as an array for multiallelic sites. This should work if you first run SelectVariants with "-restrictAllelesTo BIALLELIC".

    Also, you almost certainly don't want to run UG in -U mode - either fix your input files or drop that option.

Answers

Sign In or Register to comment.