MRNM problem for the .sam output file of tophat

[yjlui]

I have the same problem too, and here are several examples from my SAM file.

GAII04_0001:1:118:9134:2400#0 147 chr1 14092 255 36M = 14197 0 GAGCAAACTCCAAGACACCTTCTACCCCGACACCAG CCCCCCDCCCCCCCCCCCCCCCC@CCCCCCCCCCCC NM:i:2
GAII04_0001:1:118:9134:2400#0 99 chr9 14197 255 36M = 14092 0 AACCACTTGAGCAAACTCCAAGACACCTTCTACCCC CCCCCCCCCCCCCCA=CCCCDCDCCCC@CCCBBDCD NM:i:1
GAII04_0001:1:119:10145:4447#0 163 chr1 18372 255 36M = 18561 0 GGGAGATCCTGCCATGGAGAAGATCACAGAGGCTGG CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC NM:i:0
GAII04_0001:1:119:10145:4447#0 83 chr9 18561 255 36M = 18372 0 AGAGCTGCCAGGCCAGGCCCTCAGGCAAGGGCTCTG DC?CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC NM:i:0
GAII04_0001:1:17:2127:14575#0 147 chr19 40331067 255 36M = 40331078 0 TGGCTCCACCATCACATTCTTCCCCGACTGGTTTCC CCCCDCCDCCCCCDCCCCCCCCCCCCCCCCCCCCCC NM:i:1
GAII04_0001:1:17:2127:14575#0 99 chr15 40331078 255 36M = 40331067 0 CTTCTTCAAGGTGATATAGACGCTGCCCGACGTCCG CCCCCCCCCCCBCCCCCCCCCCCCCCCCCCCCDCCC NM:i:0
GAII04_0001:1:94:17511:16302#0 147 chr6 42048427 255 36M = 42048548 0 CTCTTATAATTTTGTTCTGAGGGGAGACAGGGAAAG CCCCCCCCBBCCCCCCCCCCCCCCCCCCCCCCCCCC NM:i:0
GAII04_0001:1:94:17511:16302#0 99 chr1 42048548 255 36M = 42048427 0 TCCCTTTTGTTTTCAAACCCTTCTTGGTCTTTTTGG CCCCCCCCCCBBBBCCCCCCCCCCCCCCCCCCCBBC NM:i:1

For some reason, most of the reads that have this error are mapped to chr1 and chr9.

The commands that I used to run TopHat and the corresponding reads in the sequence files are as follows:

tophat --solexa1.3-quals -p 4 -r 170 /databank/indices/bowtie/GRCh37/GRCh37 s_1_1_sequence.txt s_1_2_sequence.txt
--
@GAII04_0001:1:118:9134:2400#0/1
AACCACTTGAGCAAACTCCAAGACACCTTCTACCCC
+GAII04_0001:1:118:9134:2400#0/1
bbbbbbbbbbbbbb`\bbbbcbcbbbb_bbbaacbc
@GAII04_0001:1:119:10145:4447#0/1
CAGAGCCCTTGCCTGAGGGCCTGGCCTGGCAGCTCT
+GAII04_0001:1:119:10145:4447#0/1
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb^bc
@GAII04_0001:1:17:2127:14575#0/1
CTTCTTCAAGGTGATATAGACGCTGCCCGACGTCCG
+GAII04_0001:1:17:2127:14575#0/1
bbbbbbbbbbbabbbbbbbbbbbbbbbbbbbbcbbb
@GAII04_0001:1:94:17511:16302#0/1
TCCCTTTTGTTTTCAAACCCTTCTTGGTCTTTTTGG
+GAII04_0001:1:94:17511:16302#0/1
bbbbbbbbbbaaaabbbbbbbbbbbbbbbbbbbaab
--
@GAII04_0001:1:118:9134:2400#0/2
CTGGTGTCGGGGTAGAAGGTGTCTTGGAGTTTGCTC
+GAII04_0001:1:118:9134:2400#0/2
bbbbbbbbbbbb_bbbbbbbbbbbbbbbbcbbbbbb
@GAII04_0001:1:119:10145:4447#0/2
GGGAGATCCTGCCATGGAGAAGATCACAGAGGCTGG
+GAII04_0001:1:119:10145:4447#0/2
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
@GAII04_0001:1:17:2127:14575#0/2
GGAAACCAGTCGGGGAAGAATGTGATGGTGGAGCCA
+GAII04_0001:1:17:2127:14575#0/2
bbbbbbbbbbbbbbbbbbbbbbcbbbbbcbbcbbbb
@GAII04_0001:1:94:17511:16302#0/2
CTTTCCCTGTCTCCCCTCAGAACAAAATTATAAGAG
+GAII04_0001:1:94:17511:16302#0/2
bbbbbbbbbbbbbbbbbbbbbbbbbbaabbbbbbbb

[Bio.X2Y]

Hi,

I was wondering if this is a recognised bug?

Well actually two questions:

(1) should TopHat be mapping mate pairs to separate chromosomes in the first place?
(2) if this is allowed, should the MRNM field have a different value?

Thanks for your time.

[Gangcai]

I have send you the email, please check it~~

[Cole Trapnell]

Can you construct a small dataset I can use to reproduce this issue? It's probably a bug in matching up the left and right reads from each pair, and I haven't seen it before. If you email me a test set, I'll run it through the debugger. Thanks for your feedback.

[Gangcai]

Hi everyone,
MRNM is the 7th field in the sam file. I got the sam output from tophat, it seems that the MRNM is not correct in my case.

$ grep "HWI-EAS244_1_1_116_848_845_0_1_12496949" accepted_hits.sam
HWI-EAS244_1_1_116_848_845_0_1_12496949;2 163 chr1 14628 255 75M = 14859 0 CCTGGCTGTGTCCATGTCAGAGCAACGGCCCAAGTCTGGGTCTGGGGGGGCAGGTGTCATGGAGCCCCCTACGCT ABA?BBBB=A>BABA@2@B=;;=9<AB8ABB38=0>A?<=4AA<9??=########################### NM:i:2
HWI-EAS244_1_1_116_848_845_0_1_12496949;1 83 chr9 14859 255 76M = 14628 0 GGCAAAGGAGGGATGGAGTCTGACACGCGGACAAAGGCTCCTCCGGGCCCCTCACCAGCCCCAGGTCCTTTCCCAG 7>@@)@B=1@B;<(8@?@@9@A>B@BBABAAA@AAAB@@A@<AABABBBA?BB@BBBBB@BABCBCCCCCCCCCCB NM:i:0


The MRNM is '=' here, but one is on chr1 and another is on chr9.

From the manual , if MRNM is '=', the mate pairs should be on the same reference(same chromosome), but here it's on different chromosome. Does anybody meet the same problem? How to solve this problem?

version of tophat: tophat1.0.13
part of the shell code for tophat running is:
"
tophat -r 200 --mate-std-dev 200 -a 8 -m 0 -I 1000000 -p 4 -g 100 -o $OUTDIR/ --solexa1.3-quals --coverage-search --microexon-search --max-segment-intron 1000000 -G $GFF3 $INDEXDIR/hg19 $READDIR/s_1.fq $READDIR/s_2.fq
"

By the way, the FLAG (second field) in the .sam file is not correct in this case:
163="10100011"
83="1010011"
This means "The read is mapped in a proper pair", but they are on the different chromosome, how can they be properly paired?