discordant alignments when bowtie2 set to --no-discordant

Question

[mark.rose]

Hi All

Can anyone explain why I get many paired-end alignments in the wrong orientation and 17,000 bp apart with the following command

bowtie2 --local -N 0 -L 25 --threads 10 -I 0 -X 750 --no-discordant -x reference.fa -1 F05-63_TGGAACAA_L001_R1_001.sample.fq -2 F05-63_TGGAACAA_L001_R2_001.sample.fq -S F05_63-reads_1and2-vs-reference.fa.sam

Thanks
Mark

[Ashutosh Pandey]

Can you print a few of such alignments from the SAM output that you got?

[mark.rose]

Here's an example

M00831:170:000000000-ABEBU:1:2118:23675:13437   113     22502   194     44      250M    =       17204   0       AACCCGGGGGCGCGCCGGGACCCCTGCAGTGCAGCGTGACCCGGTCGTGCCCCTCTCTAGAGATAATGAGCATTGCATGTCTAAGTTATAAAAAATTACCACATATTTTTTTTGTCACACTTGTTTGAAGTGCAGTTTATCTATCTTTATACATATATTTAAACTTTACTCTACGAATAATATAATCTATAGTACTACAATAATATCAGTGTTTTAGAGAATCATATAAATGAACAGTTAGACATGGTCT  #######C3GGDDCGC?4:C>GFE@*GF<75@+,;,:8:EEE===>8C8+@>6E@9;;DE;@=6,=@9DCGE,FGF=?:F;A;;AGFDGCGDF9GGFE<;:9FF8FC+D4FGGGGGFAF<8FE@<AGFCEE<FEA,GFF9FGFGCF9CDGAGGGGFFEF<GGGGGGGGGCGGFGGGDFGGFGGFCCCGGGGGFFEDFGDDGGF9GGGGGFC<DFFEGFFCGFGFCDGGFEEAFC9GFF@@<FGGCCCCCC  AS:i:500        XN:i:0  XM:i:0  XO:i:0  XG:i:0  NM:i:0MD:Z:250        YT:Z:UP
M00831:170:000000000-ABEBU:1:2118:23675:13437   177     22502   17204   22      4S96M150S       =       194     0       TTTACAGTAGTGCCGCGTGATCCTTTGGCAACTAGCTAGATTAATTAACGCAATCTGTTATTGAGTTGTCTAAGCGTCAATGTGTTTACACCACAAAATACACAGCGGCCTCGCAGGCACAGGCACGCGAGGGCACACAACTCCACTTTACAAGCCTTTACACATACACAAGCGAGGGCACCACGCCCTACATCGGCTCAACCGTAGGAATAATCCCCATCTCTCTATAATTTAGTTACATTATCCTAAG  ################################################EA92@,,C9FE@@44,,5@,3,+4+:+,9;7,,,@7,3,36,+,673@;DF=A>3@8+6+B@++:<E,GD?9,++44+@4,,EE<=A,GFGFDEC,,,9B,E4,?F<EF<C,96CC,6@6:,,:CEFC:,,C@66+@FE<E6@+@@@8GF7,,:B,6;,C6,,C,C8,,,E8GFC9EDFE,AC,,,,@,FAE;,;,CCCCCC  AS:i:156        XN:i:0  XM:i:8  XO:i:0  XG:i:0NM:i:8  MD:Z:3G1T0C9G5A35A18T14T3       YT:Z:UP

Answer 1

[Devon Ryan]

You need the --no-mixed option. You're not the only person to have found this a bit weird :)

[mark.rose]

So you're saying that I have to disallow orphaned alignments if I want to disallow discordant alignments? This would be less than optimal.

[Devon Ryan]

Correct, since there's no distinction made between singletons from a pair that both map and singletons from a pair where only one maps. The only other option is to allow all of this and just filter the output with a small script.