IndelRealigner ERROR

[mboursnell]

Whe I run IndelRealigner I get an ERROR like this. It is a whole screen of chrUn something, ending like this:

chrUn_AAEX03026059, chrUn_AAEX03026060, chrUn_AAEX03026061, chrUn_AAEX03026062, chrUn_AAEX03026063, chrUn_AAEX03026064, chrUn_AAEX03026065, chrUn_AAEX03026066, chrUn_AAEX03026067, chrUn_AAEX03026068, chrUn_AAEX03026069, chrUn_AAEX03026070, chrUn_AAEX03026071, chrUn_AAEX03026072, chrM]
##### ERROR ------------------------------------------------------------------------------------------

The command is this:

java -Xmx8g -Djava.io.tmpdir=javatempdir -jar GenomeAnalysisTK.jar -T IndelRealigner -R canfam3.fasta -I ABC.realigned.rg.bam -targetIntervals ABC.bam.intervals -o ABC.clean.bam -known dummy_DBSNP.vcf -compress 0 -model USE_READS


Any ideas? It all works OK with canfam2 so it seems to be something to do with the canfam3 REF sequence.

[DZhang]

Hi mboursnell,

Please check if "chrUN..." and "chrM" are in your reference sequence used in mapping.

Best regards,
Douglas

https://www.contigexpress.com

[mboursnell]

Yes they both are. chrM only once.

I think now it's down to the the dummy DBSNP file; I don't what to filter any SNPs but it needs file so I have to make a dummy one. Not sure how to make one that is innocuous...

[DZhang]

In that case, you do not have to supply -known as it is optional.

Best regards,
Douglas

https://www.contigexpress.com

[mboursnell]

You're right! I'm sure it used to be required. Maybe that was with UnifiedGenotyper.

[DZhang]

I believe it helps improve the results if you have a list of known indels.

Best regards,
Douglas

https://www.contigexpress.com