SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
samtools calmd problem dgtnk Bioinformatics 1 11-29-2011 01:57 AM
samtools calmd -e nergaon Bioinformatics 1 05-29-2011 05:23 AM
samtools colorspace calmd is incredibly slow najoshi Bioinformatics 1 03-31-2011 07:33 PM
samtools calmd -- realignment? brentp Bioinformatics 3 02-14-2011 10:12 AM
samtools calmd -r a11msp Bioinformatics 0 11-17-2010 09:48 AM

Reply
 
Thread Tools
Old 01-11-2012, 03:58 PM   #1
adeirossi
Member
 
Location: California

Join Date: Jan 2012
Posts: 11
Default samtools: BAQ differs between mpileup and calmd

Hello,

Though samtools is a wonderful tool, the other day I came across a behavior that left me confused. I have found some reads in an Illumina dataset where the BAQ calculation appears to give a different result depending on whether it is performed by samtools mpileup or samtools calmd. This contradicts my assumption that BAQ-adjusted base qualities would be the same regardless of whether mpileup or calmd is used. (I am running samtools version 0.1.17 on Linux.)

To demonstrate, I put one such read into its own BAM file and ran mpileup and calmd:

$ samtools view my.bam
SEQUENCER02_165:1:1207:8548:44865 99 chr3.fa 178936033 254 50M = 178936190 207 AAATGACAAAGAACAGCTCAAAGCAATTTCTACACGAGATCCTCTCTCTG CCCFFFFFHHHHHJJJJJJIJJJIJJJJJJJJJJJJJJJJIJIJJJJJJJ BC:Z:CGATGT XD:Z:50 SM:i:3 AS:i:406
## Here's the original base quality string.

$ samtools calmd -Ar my.bam hg19_sorted.bam.fa
@HD VN:1.0 SO:coordinate
...
SEQUENCER02_165:1:1207:8548:44865 99 chr3.fa 178936033 254 50M = 178936190 207 AAATGACAAAGAACAGCTCAAAGCAATTTCTACACGAGATCCTCTCTCTG ;;>FFFFFHHHHHJJJJJJIJJJIJJJJJJJJJJJJJJJJIJIJJJJJJE BC:Z:CGATGT XD:Z:50 SM:i:3 AS:i:406 ZQ:Z:[email protected]@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@E NM:i:0 MD:Z:50
## Here's the BAQ-adjusted base quality string, using calmd.

$ samtools mpileup -A -Q 0 -d 10000000 -f hg19_sorted.bam.fa my.bam
[mpileup] 1 samples in 1 input files
(mpileup) Max depth is above 1M. Potential memory hog!
chr3.fa 178936033 A 1 ^~. C
chr3.fa 178936034 A 1 . C
chr3.fa 178936035 A 1 . C
chr3.fa 178936036 T 1 . F
chr3.fa 178936037 G 1 . F
chr3.fa 178936038 A 1 . F
chr3.fa 178936039 C 1 . F
chr3.fa 178936040 A 1 . F
chr3.fa 178936041 A 1 . H
chr3.fa 178936042 A 1 . H
chr3.fa 178936043 G 1 . H
chr3.fa 178936044 A 1 . H
chr3.fa 178936045 A 1 . H
chr3.fa 178936046 C 1 . J
chr3.fa 178936047 A 1 . J
chr3.fa 178936048 G 1 . J
chr3.fa 178936049 C 1 . J
chr3.fa 178936050 T 1 . J
chr3.fa 178936051 C 1 . J
chr3.fa 178936052 A 1 . I
chr3.fa 178936053 A 1 . J
chr3.fa 178936054 A 1 . J
chr3.fa 178936055 G 1 . J
chr3.fa 178936056 C 1 . I
chr3.fa 178936057 A 1 . J
chr3.fa 178936058 A 1 . J
chr3.fa 178936059 T 1 . J
chr3.fa 178936060 T 1 . J
chr3.fa 178936061 T 1 . J
chr3.fa 178936062 C 1 . J
chr3.fa 178936063 T 1 . J
chr3.fa 178936064 A 1 . J
chr3.fa 178936065 C 1 . J
chr3.fa 178936066 A 1 . J
chr3.fa 178936067 C 1 . J
chr3.fa 178936068 G 1 . J
chr3.fa 178936069 A 1 . J
chr3.fa 178936070 G 1 . J
chr3.fa 178936071 A 1 . J
chr3.fa 178936072 T 1 . J
chr3.fa 178936073 C 1 . I
chr3.fa 178936074 C 1 . J
chr3.fa 178936075 T 1 . I
chr3.fa 178936076 C 1 . J
chr3.fa 178936077 T 1 . J
chr3.fa 178936078 C 1 . J
chr3.fa 178936079 T 1 . J
chr3.fa 178936080 C 1 . J
chr3.fa 178936081 T 1 . J
chr3.fa 178936082 G 1 .$ J
## Here's the BAQ-adjusted base quality string, using mpileup.
## Note that calmd and mpileup give different BAQ-adjusted base qualities at the start and end of the read.



None of the following modifications to the options seems to make any difference in this case:
  • omitting the -A option with calmd
  • using the default -Q value with mpileup
  • using the -E option with mpileup

Why do the BAQ-adjusted base qualities differ between mpileup and calmd? Any thoughts or explanation would be greatly appreciated. Thanks in advance!
adeirossi is offline   Reply With Quote
Old 04-12-2018, 02:57 PM   #2
kjohnsen
Junior Member
 
Location: Utah, USA

Join Date: Apr 2018
Posts: 1
Default

I haven't conducted as thorough an investigation of the matter as adeirossi, but six years later I can confirm that in the pipeline we're developing, using calmd -bAr beforehand is not yielding the same results as performing the BAQ calculation in mpileup.
kjohnsen is offline   Reply With Quote
Reply

Tags
samtools

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 03:01 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO