I recieved two already aligned to hg19 and filtered BAM files and wanted to proceed with Variant Calling on VarScan Somatic for comparing normal and diseased samples. I came across the following error message:
[W::hts_idx_load2] The index file is older than the data file: normal.bam.bai
[E::mpileup] fail to parse region ‘chr11_gl000202_random:’ with normal.bam
Is there something I can do to solve this problem and use these files in VarScan or should I restart the quality control and mapping with the respective FastQ files? Although there is a workflow for that on Galaxy Training, I´d rather use the BAM files because I have never mapped data before, I usually work with BAM and VCF files. Thanks in advance, any help would be most welcome!
Hi Anna,
first, please don’t post issues in multiple places (VarScan error: index file older than data file and fail to parse certain regions · Issue #1879 · galaxyproject/galaxy-hub · GitHub) - it won’t help anyone.
Now, in your error message E stands for error and W for warning so what causes your tool run to fail is the “chr11_”… issue. Most likely, your BAM file has been generated against a version of hg19 that contained only the canonical chromosomes, chr11 for example, but not incompletely assembled scaffolds like the one mentioned. Varscan is really strict about this and insists on identity between the reference genome sequence names and the names listed in the header of your BAM file (which you can check yourself by clicking on the eye icon of that dataset and looking for header lines starting with @SQ.
You didn’t mention where you have been running this job, but on usegalaxy.eu, there is a reference genome called “hg19 Canonical”, which is identical to the full hg19 for all the standard chromosomes, but is lacking the additional scaffolds, and which you may try as your reference.
The warning about the index file is of no relevance in Galaxy and you can safely ignore it.
Cheers,
Wolfgang
Hello Wolfgang!
First of all, I´m truly sorry for posting on both forums, I guarantee it won´t happen again!
Just finished running VarsCan with the hg19 Canonical (on galaxy.eu) and it worked perfectly. Can´t thank you enough for your help!