We've moved!
This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!

Picard AddOrReplaceReadgroups doesn't like my BAM file

kohrnbkohrnb Portland State UniversityMember
edited March 2016 in Ask the GATK team

I'm processing chloroplast-captured DNA sequence, and picard tools doesn't like a few of my files. My pipeline up to the failure point (as best I can tell) is:

Adapter-trimming with cutadapt
Quality trimming with sickle
Alignment to a reference genome (KM360047.1) using bwa aln + sampe
sorting with samtools (using -F 4 option to remove unmapped reads)
index with samtools
reheader with samtools (to make sure the file is marked as sorted)
AddOrReplaceReadGroups with picardtools

The command for AddOrReplaceReadGroups is:

java -jar /disk/scratch/Cruzan/programs/picard-tools-1.128/picard.jar AddOrReplaceReadGroups INPUT=LascalWhetSSL15_pe.sort.clean.bam OUTPUT=LascalWhetSSL15_pe.sort.rg.bam RGLB=7_2 RGPL=Illumina RGPU=TAATGCGC-GGCTCTGA RGSM=LascalWhetSSL15

and the message printed on screen is:

[Thu Mar 17 13:03:23 PDT 2016] picard.sam.AddOrReplaceReadGroups INPUT=LascalWhetSSL15_pe.sort.clean.bam OUTPUT=LascalWhetSSL15_pe.sort.rg.bam RGLB=7_2 RGPL=Illumina RGPU=TAATGCGC-GGCTCTGA RGSM=LascalWhetSSL15 RGID=1 VERBOSITY=INFO QUIET=false VALIDATION_STRINGENCY=STRICT COMPRESSION_LEVEL=5 MAX_RECORDS_IN_RAM=500000 CREATE_INDEX=false CREATE_MD5_FILE=false
[Thu Mar 17 13:03:23 PDT 2016] Executing as [email protected] on Linux 2.6.32-573.12.1.el6.x86_64 amd64; OpenJDK 64-Bit Server VM 1.7.0_91-mockbuild_2015_10_21_19_56-b00; Picard version: 1.128(c8e12338d226532b30e9ecdbf33180a073c3ffc7_1421081159) IntelDeflater
INFO 2016-03-17 13:03:23 AddOrReplaceReadGroups Created read group ID=1 PL=Illumina LB=7_2 SM=LascalWhetSSL15

[Thu Mar 17 13:03:23 PDT 2016] picard.sam.AddOrReplaceReadGroups done. Elapsed time: 0.00 minutes.
To get help, see http://broadinstitute.github.io/picard/index.html#GettingHelp
Exception in thread "main" htsjdk.samtools.SAMFormatException: SAM validation error: ERROR: Record 1, Read name D00735:83:C8TJUANXX:4:2208:17299:69297, Mate Alignment start should != 0 because reference name != *.
at htsjdk.samtools.SAMUtils.processValidationErrors(SAMUtils.java:439)
at htsjdk.samtools.BAMFileReader$BAMFileIterator.advance(BAMFileReader.java:643)
at htsjdk.samtools.BAMFileReader$BAMFileIterator.(BAMFileReader.java:616)
at htsjdk.samtools.BAMFileReader$BAMFileIterator.(BAMFileReader.java:604)
at htsjdk.samtools.BAMFileReader.getIterator(BAMFileReader.java:312)
at htsjdk.samtools.SamReader$PrimitiveSamReaderToSamReaderAdapter.iterator(SamReader.java:424)
at htsjdk.samtools.SamReader$PrimitiveSamReaderToSamReaderAdapter.iterator(SamReader.java:334)
at picard.sam.AddOrReplaceReadGroups.doWork(AddOrReplaceReadGroups.java:113)
at picard.cmdline.CommandLineProgram.instanceMain(CommandLineProgram.java:187)
at picard.cmdline.PicardCommandLine.instanceMain(PicardCommandLine.java:95)
at picard.cmdline.PicardCommandLine.main(PicardCommandLine.java:105)

I looked at the first lines of the bam file in question, and it shows:
D00735:83:C8TJUANXX:4:1301:7783:38994 145 gi|704001763|gb|KM360047.1| 1 37 55M = 150478 150422 GGCGAACGACGGGAATTGAACCCGCGCATGGTGGATTCACAATCCACTGCCTTGA E/F0E/EB=/E111C1=0/<C//E9>[email protected]=>[email protected]=1=1GFC=/GGGFDE
D00735:83:C8TJUANXX:4:2105:1447:79915 153 gi|704001763|gb|KM360047.1| 5 37 49M1I35M = 5 0 AACGACGGGAATTGAACCCGCGCATGGTG

so the first line does indeed seem to have a position of 0 (for both self and mate; I'm guessing this has to do with how BWA handles unmapped mates, along with the fact that the chloroplast genome is circular). Any idea how I can keep this from happening in the future? I'd like to run these samples through indel-realigner, and I need AddOrReplaceReadgroups working first.

I'm using BWA version 0.7.5a-r405, Samtools version 0.1.17 (for reasons of downstream-pipline processes), and PicardTools version 1.128 (I also tried using version 1.141, which is the most recent version that will run on my cluster, but it failed the same way).

Any advice about how to avoid this problem in the future would be appreciated; I want to be able to automate this part of processing, and until I get this worked out, I can't.

Issue · Github
by Sheila

Issue Number
Last Updated
Closed By

Best Answers


  • SheilaSheila Broad InstituteMember, Broadie ✭✭✭✭✭


    Thanks for reporting the potential solution. Yes, we were thinking the circular reference might be the issue, as GATK tools are not designed to handle it.

    Let us know how everything works!


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    The problem with this strategy is that for the portion that you copied, reads will map non-uniquely so they will be considered suspect. And for what does get mapped with useable scores, the coverage will be split between the two copies so calls made in those regions will probably be only weakly supported.

    I'm not sure what is the right way to deal with this problem but I don't think the solution presented here is going to work out. You should look up how others deal with this in e.g. the microbial literature.

Sign In or Register to comment.