Why bowtie report same number of reads with different - k parameters?

Question

[xiaoleiusc]

Hi, All,

I use bowtie to align my short RNA reads to the hg19 genome. I tested different ways of reporting alignment read number by changing -k parameter, however, I found bowtie report the number of aligned reads even with different -k parameters. This behavior of bowtie is also inconsistent with Qualimap tool analysis of the bam file, Qualimap seems to make more sense that -k 10 should give me more reported reads than -k 1. I attach my bowtie and Qualimap analysis below:

bowtie -f -v 2 -m 10 -k 10 -p 18 -S /miniconda3/data/hg19index/hg19_index input.fa output.sam 2>&1 | tee output.log

reads processed: 15013

reads with at least one reported alignment: 5411 (36.04%)

reads that failed to align: 4642 (30.92%)

reads with alignments suppressed due to -m: 4960 (33.04%)

Reported 17093 alignments to 1 output stream(s)

bowtie -f -v 2 -m 10 -k 5 -p 18 -S /miniconda3/data/hg19index/hg19_index input.fa output.sam 2>&1 | tee output.log

reads processed: 15013

reads with at least one reported alignment: 5411 (36.04%)

reads that failed to align: 4642 (30.92%)

reads with alignments suppressed due to -m: 4960 (33.04%)

Reported 13295 alignments to 1 output stream(s)

bowtie -f -v 2 -m 10 -k 1 -p 18 -S /miniconda3/data/hg19index/hg19_index input.fa output.sam 2>&1 | tee output.log

reads processed: 15013

reads with at least one reported alignment: 5411 (36.04%)

reads that failed to align: 4642 (30.92%)

reads with alignments suppressed due to -m: 4960 (33.04%)

Reported 5411 alignments to 1 output stream(s)

Qualimap bam analysis results:

-m 10 - k 10 Qualimap results:Number of reads 26,695 Mapped reads 17,093 / 64.03% ; Unmapped reads 9,602 / 35.97%;

-m 10 - k 5 Qualimap results: Number of reads 22,897 Mapped reads 13,295 / 58.06% ; Unmapped reads 9,602 / 41.94%;

-m10 -k 1 Qualimap results:Number of reads 15,013 Mapped reads 5,411 / 36.04% ; Unmapped reads 9,602 / 63.96%

Any inputs are appreciated.

Xiao

[ATpoint]

You may consider using the code option to highlight your code examples for better visual appearance.

[xiaoleiusc]

thanks. Just did, works great!!

[wm]

First of all, be noticed, reads and alignments are different.

1 read, could have multiple alignments reported in the sam file.

So, the number of reads are identical with the same parameter-v 2 -m 10.

But, thealignments will not, using different -k. (you can check it by samtools view -c out.sam),

-k 10 is supposed to report more alignments than -k 1.