anomalous alignment with bowtie2

Question

[mark.rose]

Hi All

I getting what appears to be an anomalous alignment with bowtie2 (v2.3.4). Consider the following:

reference1

M00831:461:000000000-CP4BP:1:2105:10688:18697 99  AGRO_LBA4404_1|NODE_23  996 31  100M  = 1213  267 CTCGACTGGCAATGAGAAGTTGCTCGCGCGATAGAACGTCGCGGGGTTTCTCTAAAAACGCGAGGAGAAGATTGAACTCACCTGCCGTAAGTTTCACCTC  CCCCCGGGGGGGGGGGGGGGGGGFFGGGGEGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGFGFGGGGGGGGGGGGGGGGGFGGGGFGGG  AS:i:200  XS:i:200  XN:i:0  XM:i:0  XO:i:0  XG:i:0  NM:i:0  MD:Z:100  YS:i:100  YT:Z:CP

reference 2

M00831:461:000000000-CP4BP:1:2105:10688:18697   83      24653-1agrWset1 11508   44      100M    =       11341   -267    GAGGTGAAACTTACGGCAGGTGAGTTCAATCTTCTCCTCGCGTTTTTAGAGAAACCCCGCGACGTTCTATCGCGCGAGCAACTTCTCATTGCCAGTCGAG    GGGFGGGGFGGGGGGGGGGGGGGGGGFGFGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGEGGGGFFGGGGGGGGGGGGGGGGGGCCCCC    AS:i:200        XN:i:0  XM:i:0  XO:i:0  XG:i:0  NM:i:0  MD:Z:100

The difference in orientation aside, the only difference to my eye beween this read's alignment to two different data bases is the presence of the XS tag in the alignment to reference 1, an indication that the read aligns to more than one location in that reference. In other words it seems there is no difference in the quality of the alignment itself, as indicated by the AS and MD tags.

However, when I align this read with blast there is a difference in the alignments as show below. (sorry, I can't seem to get it to maintain the formatting for the alignments)

Database: reference 1    
Query= M00831:461:000000000-CP4BP:1:2105:10688:18697
Length=100
                                                                      Score     E
Sequences producing significant alignments:                          (Bits)  Value
  seq1    185     1e-47

> seq1
Length=176344

 Score = 185 bits (100),  Expect = 1e-47
 Identities = 100/100 (100%), Gaps = 0/100 (0%)
 Strand=Plus/Plus

Query  1       ACGGATAAAGTTGTTGCACTCGAGCTAGGAGCAAGTGATTTTATCGCTAAGCCGTTTAGT  60
               ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Sbjct  123352  ACGGATAAAGTTGTTGCACTCGAGCTAGGAGCAAGTGATTTTATCGCTAAGCCGTTTAGT  123411

Query  61      ACGAGAGAGTTTCTTGCACGCATTCGGGTTGCCTTGCGCG  100
               ||||||||||||||||||||||||||||||||||||||||
Sbjct  123412  ACGAGAGAGTTTCTTGCACGCATTCGGGTTGCCTTGCGCG  123451

Database: reference 2
Query= M00831:461:000000000-CP4BP:1:2105:10688:18697
Length=100
                                                                      Score     E
Sequences producing significant alignments:                          (Bits)  Value
  seq1                                                     163     7e-44

> seq1
Length=14823

 Score = 163 bits (88),  Expect = 7e-44
 Identities = 96/100 (96%), Gaps = 0/100 (0%)
 Strand=Plus/Plus

Query  1      ACGGATAAAGTTGTTGCACTCGAGCTAGGAGCAAGTGATTTTATCGCTAAGCCGTTTAGT  60
              |||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||
Sbjct  11307  ACGGATAAAGTTGTTGCACTCGAGCTAGGAGCAAGTGATTTTATCGCTAAGCCGTTCAGT  11366

Query  61     ACGAGAGAGTTTCTTGCACGCATTCGGGTTGCCTTGCGCG  100
              |  ||||||||||| |||||||||||||||||||||||||
Sbjct  11367  ATCAGAGAGTTTCTAGCACGCATTCGGGTTGCCTTGCGCG  11406

In the blast result there is a 4 mismatch difference between the alignments of this read to reference 1 and 2, despite bowtie2 indicating both alignments are are equal and perfect.

Thanks for your help

Mark

[GenoMax]

Are the blast alignments against two different original references (which appear to be of different length based on info above)? Are you selecting only one top alignment? Are there other alignments reported?

[mark.rose]

The blast alignments are against to different blast dbs.

For reference 1 I am selecting the top hit (so that bowtie2 produces an XS tag for the reference 1 alignment seems correct). There is only 1 sequence in the blast db for reference 2.

[Asaf]

The query sequences in the BLAST are not the same as in the sam file, did I miss something?