It looks like you're new here. If you want to get involved, click one of these buttons!
I am running the INDEL model of VariantRecalibrator with GATK 1.6 (last license-compatible version in our hands), and the very last step seems to crash when writing the plot using plot_Tranches.R (see below).
All the previous steps seem to run well on 9G RAM and I presume this plotting step just produces an image to manually inspect and nothing else. Is it possible to switch off the plotting step at all?
java -Xmx8g -jar /my/path/GATK/GenomeAnalysisTK.jar -et NO_ET -T VariantRecalibrator -R /my/path/WholeGenomeFASTA/genome.fa -input /my/path/genome.vcf.tmp2 -resource:mills,known=true,training=true,truth=true,prior=12.0 /my/path/Annotation/Mills_Devine_2hit.indels.hg19.vcf -an QD -an FS -an ReadPosRankSum -an DP --mode INDEL -recalFile /my/path/test/file.INDEL.recal -tranchesFile /my/path/test/file.INDEL.tranches
[...] INFO 10:46:37,310 VariantRecalibrator - Writing out recalibration table... INFO 10:46:46,202 VariantRecalibrator - Executing: Rscript (resource)org/broadinstitute/sting/gatk/walkers/variantrecalibration/plot_Tranches.R /my/path/test/file.INDEL.tranches 2.15
PS: I tried adding the option "-rscriptFile /dev/null", but it also seems to crash at the same step: INFO 11:25:10,870 VariantRecalibrator - Executing: Rscript /dev/null