bwa sampe: proper pair but on different contigs!!??!!

[GoneSouth]

Dear all,


Does anyone have an idea how the following is possible:
I have reads mapped in a proper pair (as indicated by the sam-flag) but they map to different contigs!!!???

HWUSI-EAS300R:7:1:15:1404#0 147 FW_DM_LINE_Jockey 128 29 74M FW3_DM_LINE_Jockey 3131 0 TGCAAGATCGCTTAAATACATAGTGAATTGTTATCTTAAATAATAAAACTATGAGTCAGAATGACACTCGCGCC Y^S[]^\[]a_XSZ[_]]_`_`]```_^a^`^`[aa__`]V]```aa\a_`]aaaaaaaa`Ta\a`aaaba`aa XT:A:U NM:i:0 SM:i:29 AM:i:29 X0:i:1 X1:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:74
HWUSI-EAS300R:7:1:15:1495#0 147 Gypsy4_LTR_LTR_Gypsy 112 60 74M Gypsy4_I_LTR_Gypsy 6216 0 CATTCCACTGCCCGGAGCGTGTGAAGCGCAATGTCAGCATTCTGCCGTGAGCGCTGCTTCAAAAGACGGGCTAC XUPM^NHLSMW\SWSPM\MW]PW\TZ\aPMP^MS^S]]Z^M_^X]^Z^]Z^]`a]^Z_\aaS]Z`Sa]a`_a\a XT:A:U NM:i:3 XN:i:1 SM:i:37 AM:i:37 X0:i:1 X1:i:0 XM:i:3 XO:i:0 XG:i:0 MD:Z:5T32C22G12
HWUSI-EAS300R:7:1:22:1504#0 147 FW_DM_LINE_Jockey 85 29 74M FW3_DM_LINE_Jockey 3125 0 AACTAAATAAAAAATCTGAAAGCGAAAGAGACGCTCTATGCGATGCAAGATCGCTTAAATACATAGTGAATTGT ]N^I_^WG[[[_YNFQP[XGM\_^^S\a__^``_Y[a^\_a_```aaa`a]a`a````ba_baa`a_bbaabaa XT:A:U NM:i:0 SM:i:29 AM:i:29 X0:i:1 X1:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:74
HWUSI-EAS300R:7:1:25:1975#0 83 BLOOD_I_LTR_Gypsy 145 29 13M3D61M BLASTOPIA_LTR_LTR_Gypsy 271 0

Hope anyone can help on this!!
best ro

[maubp]

Could be a bug in the mapping tool used. What tool and what version was it?

[GoneSouth]

Mapper: bwa
Version: 0.57
command bwa aln -n 0.01 -o 2 -e 12 -d 12 -t 2 etc

[maubp]

Is there any obvious link between the contigs, in particular are they subsequent entries in the FASTA reference file?

[Lee Sam]

I was under the impression that BWA concatenates all the references together and aligns reads against that long string. Might it have something to do with that?

[GoneSouth]

Yes they are subsequent entries in the fasta file! It is the insert of a LTR transposon followed by the LTR, i.e.: this sequences are frequently found in exactly this order in the different species.
This could be an explanation for the problem than. If BWA is concatenating the sequences and measuring the distance between the mates, than it finds the difference is correct, while ignoring the fact that a contig boundary is crossed, and thus assigns the flag mapped in a proper pair.

[maubp]

Yes they are subsequent entries in the fasta file!

Given Lee Sam's post you can probably see why I asked that

i.e. This is probably a bug in BWA, wrongly marking the reads as "properly paired".

[GoneSouth]

Yes I do, many thanks for all your help!!
Now that I know whats going on I can handle this in my sam parser.
And maybee the people from Sanger will find some time to fix this in one of the next versions - I will send a bug report.
thanks ro