Mapping Issues with Bowtie2

Question

[Varun Gupta]

Hi Everyone,

I have a chimeric seq read data set, where in a read, one part of the read belongs to one region in the genome and other part belongs to the other region. I am using BOWTIE2 for mapping this read data set in local mode so that later on I can map the soft clipped part separately.

I am selecting seed length of 8 (-L 8), and I want to have 1 mismatch in it. I specified my parameters like this

Case1

bowtie2 -a -L 8 -N 1 -p 12 --local -x /user/index -U test.fastq -S test_L8_N1.sam

I get this as output

45728414 reads; of these:
  45728414 (100.00%) were unpaired; of these:
    34402970 (75.23%) aligned 0 times
    7019150 (15.35%) aligned exactly 1 time
    4306294 (9.42%) aligned >1 times
24.77% overall alignment rate

Then I used this:

Case2

bowtie2 -a -L 8 -p 12 --local -x /user/index -U test.fastq -S test_L8_N0.sam (-N 0 is the default in local mode)

I get this as output

45728414 (100.00%) were unpaired; of these:
    30962101 (67.71%) aligned 0 times
    8461008 (18.50%) aligned exactly 1 time
    6305305 (13.79%) aligned >1 times
32.29% overall alignment rate

As you see the alignment percentage gets increased. Also I looked for XM:i:2 in case 1 and found only 1 XM:i:2 read, but in case2 a lot of them. Why?

I want to allow one mismatch in seed region (-L 8) and then other mismatches in the total alignment of the read, but I am confused by the 2 results above?

Any explanation as to why this result?

Regards
Varun

[mark.ziemann]

Sorry I can't help specifically with bowtie2 parameters, but perhaps your job would be better suited to Subread, which has in-built fusion detection capability.

Detect fusions in genomic DNA sequencing data:

subread-align --reportFusions -d 50 -D 600 -i my_index -r reads1.txt -R reads2.txt 
-o subread_results.sam

[michaela_boell]

It should be exactly the other way around, are sure you did not mix up which one is which?