bwa mem -T (alignment score) not doing anything
I'm trying to control which reads are mapped to a reference genome by adjusting the the -T parameter in bwa mem, which is the minimum acceptable alignment score. The only problem is that reads with an alignment score below the threshold are being mapped anyway. I can see this when using samtools view to visualize the bam file, reads with an AS value less than the threshold are there. My data is paired end, so I tried doubling the threshold to account for both reads, but that didn't work either.
What's going on here? Do some settings in bwa mem interfere with -T?
• 4,434 views
•
link
updated
by
Istvan Albert
•
written
by
chris.bird •
•
written
by
chris.bird •
0 answers
No answers yet.
Log in to answer this question.
More posts like this
-
Mapping of paired-end ddRADseq results in 0.00% of reads pairing
written by Brendan •
-
Get the coverage after read alignment with bwa-mem2
written by maxime.policarpo
-
Fastq screen: use bwa aln and not bwa mem
written by Matteo •
-
difference between mapq=0 and samtools flag 0x900
written by bitpir
-
Different alignments rates with bwa mem (0%) and bowtie2 (82%)
written by James Ashmore
-
Multiple Hits and Unique Mapped Reads
written by fusion.slope
-
BWA mem how to know if a read is mapped uniquely?
written by steven.lin.pur •
-
DNA-seq PE150 BWA alignment
written by zhuozhu132 •
-
BWA MEM -T option. What does really do and how?
written by cpcantalapiedra
-
Bwa-Mem: Discriminate Between Reads Mapping Uniquely And Those Mapping In Multiple Positions
written by Fabio Marroni
For things like this, please give the version of BWA, full command line, the reference genome and an example of the problematic reads.
Will create a new post if it's not acceptable to comment with the same issue
I have the same issue, using bwa mem 0.7.13 as follows:
The -K setting is for deterministic alignment, as seen in this post, and I don't think it's interfering with -T, since it says in the bwa manual that -T only affects the output.
Example problematic reads from sam file:
I appreciate any insights on this!
Out of curiosity do these reads have other alignments that satisfy the
-Tcut-off elsewhere in the file?No, these reads don't have other alignments at all in the file.
I have the same issue. I set -T to 75 and I still see reads with AS:i:30, for example
It seems -T parameter acts upon the MD:Z field, which apparently indicates all matches/mismatches/indels (not as straightforward as AS:i). Unmapped reads still remain in the SAM file. In my case, however, when I convert SAM to BAM, sort it and then filter it with $ samtools view -bq 1 sorted.bam > filtered.bam, then I see the min. value for the AS:i field as I set it with -T.
perhaps what is happening is that the mapping quality is set to 0 when the score is under the threshold, so the filtering takes place with the -q 1 step