Hi,
I've tried my best to find the problem but without luck, so my first post...
First the background:
I have used HaplotypeCaller to make GVCFs per individual for 2 cohorts of 270 and 300 individuals each.
I then used CombineGVCFs to merged all the individuals in each cohort across 10Mb chunks across the genome.
These steps appear to have worked OK! Then I have been attempting to call genotypes from pairs GVCFs (one from each cohort) per region using GenotypeGVCFs with a command such as:
java -jar -Xmx6G /gpfs0/apps/well/gatk/3.2-2/GenomeAnalysisTK.jar \ -T GenotypeGVCFs \ -R ../hs37d5/hs37d5.fa \ --variant /well/hill/kate/exomes//haploc/cap270/cap270_1:230000000-240000000_merged_gvcf.vcf \ --variant /well/hill/kate/exomes//haploc/sepsis300/sepsis300_1:230000000-240000000_merged_gvcf.vcf \ -o /well/hill/kate/exomes//haploc/cap270-sepsis300/cap270-sepsis300_1:230000000-240000000_raw.vcf
This seems to start fine (and works perfectly to completion for some regions), but then throws a error for some of the files. For this example the error is:
ERROR MESSAGE: The provided VCF file is malformed at approximately line number 519534: ./.:29:75:25:0,63,945 is not a valid start position in the VCF format, for input source: /well/hill/kate/exomes/haploc/cap270/cap270_1:230000000-240000000_merged_gvcf.vcf
So to try and solve this I have grepped "./.:29:75:25:0,63,945 " from the VCF file (expecting to find a truncated line), but could find nothing wrong with the lines containing this expression. I looked carefully at the first line containing this after the last position that was written in the output - nothing out of the ordinary that I can see.
Any help/advice gratefully received! Kate