Hi,
For a PE-reads BAM file, I would like to keep only those pairs that map to a unique location. I know I can filter reads with multiple mappings with samtools (-bq 1), but this doesn't take into account paired information. I mean, if mate #1 maps uniquely to locus A, and mate #2 maps both to A and B, then we should be confident that #2 maps to A. Using samtools -bq 1 removes mate #2 in this case but I would like to keep it (because of the uniqueness of mate #1). Any idea how to solve this?
Thanks!
Federico
•
written
by
abascalfederico
1 answer
-bq 1 will not always remove multimappers, that they have a MAPQ of 0 is aligner dependent. As an example, a MAPQ of 1 can happen with multimappers produced by bowtie2.
Anyway, if you have a name-sorted file then you can write a little script with pysam to go through and filter things to whatever definition of "unique" you'd like to use.
Log in to answer this question.
-
Filter BAM to keep only alignments at an interval, ignoring reads spliced over
written by WouterDeCoster
-
Pileup of inner distance for PE reads
written by noodle
-
STAR mapping statistics don't add up with samtools count
written by Zee_S
-
count reads mapped to multiple locations only once
written by snishtala03 •
-
best way to get chimp-human homologous sites
written by abascalfederico
-
pair-end, mate-pairs coding
written by kamiljaron
-
samtools mpileup: position of each base within their reads
written by abascalfederico
-
Duplicated read names in Fastq file
written by abascalfederico
-
RNA-seq stranded protocol
written by rubic
-
Does A Bwa Bam File Contain Entries For All Unaligned Paired-End Reads?
written by Rm
