Counting Reads in BAM file per genomic position

[C9r1y]

Hello,

I am looking for a tool or script that counts the number of reads in a bam file by position, and takes into account only successfully paired reads. I would like to estimate PCR duplicates when making the (possibly erroneous, I understand!) assumption that reads with the same genomic coordinates are copies.

For example, for the following input:

Chr1:1-500
R1--------------> <---------------R2
R1--------------> <---------------R2
R1--------------> <---------------R2

Chr1:5-505
R1--------------> <---------------R2

Chr1:10-510
R1--------------> <-----------R2
R1--------------> <-----------R2
R1-------------->

I would like to output this information:
Chr1:1-500 3
Chr1:5-505 1
Chr1:10-510 2

[blancha]

bedtools multicov with the argument -p.

[C9r1y]

Thanks blancha for the response. However, I would like the position coordinates to be dictated by the read positions, not a bed file of genomic regions.

I would like to estimate the number of pcr duplicates, and will make the assumption that reads that have the exact same start and stop sites are duplicates,

[HESmith]

Then use SAMtools rmdup or Picard MarkDuplicates, since those tools identify duplicates by the same criteria you're using.

[blancha]

Sorry, tired.
How about this one? From Stanford.

http://ngsutils.org/modules/bamutils/pcrdup/

It has an option -counts.

Otherwise, you may just have to find someone to write you a custom script.
I don't know of any tool to give you exactly what you want, although it may exist.