samtools pileup c-option changes read quality

DNAjunk · 11-19-2010, 06:55 AM

Hi

I might be wrong but I got the impression that when using the c-option in the pileup process, then the read qualities are changed (near an indel, N's are placed also):
(I am using version 0.1.9 but version 0.1.10 seems to be the same). I am sure that there is an explanation, but I dont know it! Thanks for help!

Example: the differences are lines 68-71

with c option: (jump to line 67-75)
/apps/bi/samtools/c/samtools pileup -c test.bam -f test_ref.fa | less
test 67 A A 30 0 60 1 . 1
test 68 C N 0 0 0 1 . (
test 69 C N 0 0 0 1 . %
test 70 C N 0 0 0 1 . !
test 71 C N 0 0 0 1 .-1C !
test 71 * -C/-C 40 0 60 1 -C * 1 0 0 0 0
test 72 C N 0 0 0 1 * <
test 73 T T 30 0 60 1 . <
test 74 A A 30 0 60 1 . @
test 75 T T 30 0 60 1 . 1

without c option: (jump to line 67-75)
/apps/bi/samtools/c/samtools pileup test.bam -f test_ref.fa | less
test 67 A 1 . 1
test 68 C 1 . 1
test 69 C 1 . 1
test 70 C 1 . 1
test 71 C 1 .-1C 1
test 72 C 1 * <
test 73 T 1 . <
test 74 A 1 . @
test 75 T 1 . 1

Here is my test.sam file

@HD VN:0.0 GO:none SO:coordinate
@SQ SN:test LN:360
@PG ID:map2sam
GLXM4QB02GKAEH 0 test 1 255 71M1D17M1D20M1I147M * 0 0 GATCTCTTTTTCCAGTTTAAAAGTTCAGCTATTACAAGGTAAATGCAAAATGACAGTAAGGTTCCCACCCCTATTTTTCATCTTTTATGTCAAGAAGATACTTGTAAAAGAGAAGTAAGGTTAGGTATGATATTCGTTTATAAAGCATTTGGTGGAATACAGCGGTGGCAGGAAGTTGAACTGTCAACATTTTTGACCGTGTCAATTCATTTAATACATGTAGCCTGTATCTATTGAGTATATGGTACACATGCGG BAAABB55733>><94444444=99??>??<888?<<<89333=996666ABA???CDBAAAA<<;11111<@11100=?A?44444/4894443<98===993/.,,,-,559531266><866669:<<32444339@9;;<BBB???::8AAAA???DDDDDBBA?>===83../..33:8:9921,,,,,..--558875555:3...222679::9:32449898<<<<=AAAA@@<84449:666<::--

Here is the test_ref.fa file

>test
GATCTCTTTTTCCAGTTTAAAAGTTCAGCTATTACAAGGTAAATGCAAAATGACAGTAAG
GTTCCCACCCCCTATTTTTCATCTTTTATTGTCAAGAAGATACTTGCAAAGAGAAGTAAG
GTTAGGTATGATATTCGTTTATAAAGCATTTGGTGGAATACAGCGGTGGCAGGAAGTTGA
ACTGTCAACATTTTGTACCGTGTCAATTCATTTAATACATGTAGCCTGTATCTATTGAGT
ATATGGTACACATGCGGAAGTGTTCTGAGGATAAAGTCACGAAAAAAAAAAGGCACAATT
CTTGTTTTCAGGGAGATTATGTTCTAGTGAGGAGAATAATAGAAGAAGAAAGTCCCACTC

Cynthia3.14 · 08-01-2011, 01:24 AM

Hi,

I can confirm this bug, as I had also noticed base quality scores changes with -c option, when I compared the number of high quality en low quality bases of a first mapping to a remapping against a sample specific reference. It appeared that low quality (LQ) bases, in the first mapping turned into high (HQ) quality bases, at positions were initially the reads did not match the reference and did in the remapping. So, I checked this problem with 10 reads at positions where a mutation occurred:

first mapping
ref HQ_A HQ_C HQ_G HQ_T LQ_A LQ_C LQ_G LQ_T
T 0 2 0 1 0 7 0 0
second mapping
ref HQ_A HQ_C HQ_G HQ_T LQ_A LQ_C LQ_G LQ_T
C 0 9 0 1 0 0 0 0

When the pileup was made without the option -c, these changes did not happen. Also, mpileup seems to have this same problem.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
SNP quality score in Samtools pileup	wangzkai	Bioinformatics	5	09-21-2011 06:32 AM
samtools pileup base quality	swapna	Genomic Resequencing	0	02-23-2011 03:36 AM
Samtools pileup indel quality score computation	christophpale	Bioinformatics	0	07-27-2010 12:51 PM
How is SNP quality calculated in SAMTOOLS pileup?	lazyworm	Bioinformatics	2	05-28-2010 12:00 AM
get read position from Samtools pileup	rcorbett	Bioinformatics	2	02-10-2010 03:38 PM

11-19-2010, 06:55 AM	#1
DNAjunk Member Location: Western Europe Join Date: Jun 2009 Posts: 56	samtools pileup c-option changes read quality Hi I might be wrong but I got the impression that when using the c-option in the pileup process, then the read qualities are changed (near an indel, N's are placed also): (I am using version 0.1.9 but version 0.1.10 seems to be the same). I am sure that there is an explanation, but I dont know it! Thanks for help! Example: the differences are lines 68-71 with c option: (jump to line 67-75) /apps/bi/samtools/c/samtools pileup -c test.bam -f test_ref.fa \| less test 67 A A 30 0 60 1 . 1 test 68 C N 0 0 0 1 . ( test 69 C N 0 0 0 1 . % test 70 C N 0 0 0 1 . ! test 71 C N 0 0 0 1 .-1C ! test 71 * -C/-C 40 0 60 1 -C * 1 0 0 0 0 test 72 C N 0 0 0 1 * < test 73 T T 30 0 60 1 . < test 74 A A 30 0 60 1 . @ test 75 T T 30 0 60 1 . 1 without c option: (jump to line 67-75) /apps/bi/samtools/c/samtools pileup test.bam -f test_ref.fa \| less test 67 A 1 . 1 test 68 C 1 . 1 test 69 C 1 . 1 test 70 C 1 . 1 test 71 C 1 .-1C 1 test 72 C 1 * < test 73 T 1 . < test 74 A 1 . @ test 75 T 1 . 1 Here is my test.sam file @HD VN:0.0 GO:none SO:coordinate @SQ SN:test LN:360 @PG ID:map2sam GLXM4QB02GKAEH 0 test 1 255 71M1D17M1D20M1I147M * 0 0 GATCTCTTTTTCCAGTTTAAAAGTTCAGCTATTACAAGGTAAATGCAAAATGACAGTAAGGTTCCCACCCCTATTTTTCATCTTTTATGTCAAGAAGATACTTGTAAAAGAGAAGTAAGGTTAGGTATGATATTCGTTTATAAAGCATTTGGTGGAATACAGCGGTGGCAGGAAGTTGAACTGTCAACATTTTTGACCGTGTCAATTCATTTAATACATGTAGCCTGTATCTATTGAGTATATGGTACACATGCGG BAAABB55733>><94444444=99??>??<888?<<<89333=996666ABA???CDBAAAA<<;11111<@11100=?A?44444/4894443<98===993/.,,,-,559531266><866669:<<32444339@9;;<BBB???::8AAAA???DDDDDBBA?>===83../..33:8:9921,,,,,..--558875555:3...222679::9:32449898<<<<=AAAA@@<84449:666<::-- Here is the test_ref.fa file >test GATCTCTTTTTCCAGTTTAAAAGTTCAGCTATTACAAGGTAAATGCAAAATGACAGTAAG GTTCCCACCCCCTATTTTTCATCTTTTATTGTCAAGAAGATACTTGCAAAGAGAAGTAAG GTTAGGTATGATATTCGTTTATAAAGCATTTGGTGGAATACAGCGGTGGCAGGAAGTTGA ACTGTCAACATTTTGTACCGTGTCAATTCATTTAATACATGTAGCCTGTATCTATTGAGT ATATGGTACACATGCGGAAGTGTTCTGAGGATAAAGTCACGAAAAAAAAAAGGCACAATT CTTGTTTTCAGGGAGATTATGTTCTAGTGAGGAGAATAATAGAAGAAGAAAGTCCCACTC

08-01-2011, 01:24 AM	#2
Cynthia3.14 Junior Member Location: Amsterdam Join Date: Jul 2011 Posts: 1	base quality score changes option -c samtools pileup and mpileup Hi, I can confirm this bug, as I had also noticed base quality scores changes with -c option, when I compared the number of high quality en low quality bases of a first mapping to a remapping against a sample specific reference. It appeared that low quality (LQ) bases, in the first mapping turned into high (HQ) quality bases, at positions were initially the reads did not match the reference and did in the remapping. So, I checked this problem with 10 reads at positions where a mutation occurred: first mapping ref HQ_A HQ_C HQ_G HQ_T LQ_A LQ_C LQ_G LQ_T T 0 2 0 1 0 7 0 0 second mapping ref HQ_A HQ_C HQ_G HQ_T LQ_A LQ_C LQ_G LQ_T C 0 9 0 1 0 0 0 0 When the pileup was made without the option -c, these changes did not happen. Also, mpileup seems to have this same problem.