There are many blogs explaining the related post but none ended with a good solution, So I am posting with a new thread.
I want to know if there is any way to get information about number of reads uniquely mapped. Actually I found a code in a nature publication that gives number of uniquely mapped reads, but that was for single end reads. Now I am handling different projects in which some have paired end and some have single end data. In many of my data sets I get very less number of uniquely mapped reads (~10% of the total reads after removing duplicates). So to check it manually I tried counting number of unique combination of chromosome name & start positions & end positions, by converting my bam file into bed. But both the strategies are producing different numbers of uniquely mapped reads. So I want to know if there is any good and reliable way to count the uniques in any supported format. I aligned reads using bowtie2.
Just to be clear, are you talking about uniquely mapped reads or clonal reads ? When you say "
counting number of unique combination of chromosome name & start positions & end positions, ", you are talking about PCR duplicates and unique mapping is different.yes, this is the confusing part of it. what I think is removing pcr duplicates will give me all the reads that are uniquely occurring, thus this will be the number of unique. But in that protocol they have defined unique as the ones that are occurring only once, i.e they have neglected all reads that contain pcr duplicates. In this protocol I also think that if we consider only the non-pcr duplicate reads then while counting % unique we should use total as the ones without pcr duplicates (what I consider unique) not the total number of reads. That's why I want to know if there is some standard tools or some thing else that clarifies this.