Error GATK - Generating interval file

[Mandor]

Hello,

To generate interval file usign GATK is needed a .bed file to specify the positions of human exome (hg19) (let's say "all_captured_exomes.bed").

To get this file, I went to UCSC and specify certain parameters to get this .bed file. So, concretely :

http://genome.ucsc.edu/cgi-bin/hgTables?hgsid=286850271&clade=mammal&org=Human&db=hg19&hgta_group=genes&hgta_track=refGene&hgta_table=0&hgta_regionType=genome&position=chr21%3A33%2C031%2C597-33%2C041%2C570&hgta_outputType=primaryTable&hgta_outFileName=

And when I try to launch the aligner :

java -jar $GATK_HOME/GenomeAnalysisTK.jar -T RealignerTargetCreator \
-R /home/armand/mnt/cubix/Genomes/index/index/hg19/bwa_path/Homo_sapiens_assembly19.fasta \
-I $1 -log intervals.log -B:dbsnp,vcf $dbSNP \
-L ../exome_hg19_captured/all_captured_exomes.bed -o "$core_dir/${core_name}.intervals"

It gives me this error :

##### ERROR MESSAGE: File associated with name ../ref_gen_hg19/exome_hg19_captured/all_captured_exomes.bed is malformed: Interval file could not be parsed in either format. caused by Invalid command line: Contig chr1 given as location, but this contig isn't present in the Fasta sequence dictionary

Any ideas what could it be?

Thanks for your time!,

[bruce01]

Hi, you need to have the contig/chromosome IDs in the reference.fasta EXACTLY as in the VCF or BED file, and properly sorted. Also any other scaffolds/contigs in the ref.fasta not in the VCF or BED or vice-versa will cause errors. My solution was to remove all scaffolds/contigs not found in the VCF I used from my ref.fasta which has worked (so far!). Good luck!