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!

Is it possible to have a much smaller bam output from PrintReads?

I am using 3.3-0. The size of recalibrated bam is about two times the size of indel realigned bam. I tried adding the -s flag to PrintReads but the output becomes 5% smaller only.

Still, 5% smaller is still too big for me. I usually keep the recalibrated bam for igv visualization. Is it possible to make its size closer to the indel realignment bam without sacrificing the accuracy of downstream analysis and also visualization in igv???

Answers

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin

    @ymc All the ways to reduce the file size are lossy. The least lossy is getting rid of the original quals (if you kept them in the OQ tags). Otherwise, you can try quantizing your bams, or apply downsampling to coverage if there are regions with very excessive depth. You can also discard reads that fall outside of your capture intervals if you are working with exome data.

  • ymcymc Member

    Thanks for your reply. I found that -s flag removed everything except BD:Z, RG:Z and BI:Z. But in the example below, the original quality is [email protected];?8B?@BBEE8E>EEE8B>EEC8A3B?@###################### which is not the same thing as either BD:Z or BI:Z. So what is BD:Z and BI:Z??? (BTW, there is no OQ tag at all even in the output without -s)

    SRR490123.83659522 0 chr1 10332 0 28M1D5M1D42M *
    0 0 CCTAACCCTAACCCTAACCCTAACCTAACCTAACCTAACCCTAACCCTAACCCTAACCCCCAAC
    CCCTCACCCTA [email protected]><[email protected];C?DDD;A?DDC;B4?=<###########

    ##### BD:Z:OOQQRSSORQOQPMPPNPPMPPNPPPPNPPPPNPPPPNPPNPQOQQNQQOQQNOOOOOO

    OPPPPPQQQQRROOOO RG:Z:lnP3N BI:Z:SSVTSUVRVTRSTQUSQSTQUSQSTUSQSTUSRSU
    VTRTURVTRTURVTRTURSSSSSSTTTTSTTTTSTUSSSS

  • SheilaSheila Broad InstituteMember, Broadie ✭✭✭✭✭

    @ymc
    Hi,

    The BD and BI tags are for deletion and insertion qualities. The default is to not emit the original qualities (OQ tag) because they greatly increase the file size.

    -Sheila

  • ymcymc Member

    oic. Thanks for your replies. So I think it is impossible to further reduce the size for now.

Sign In or Register to comment.