Remove duplicates in featurecounts

Question

[lkmklsmn]

Hi Biostars,

I used the featureCounts function from the Rsubread package in order to count reads. The first time I used my 'regular' bam files. The second time I used Picard Tools to mark duplicates in each bam file and used this data as substrate. I am getting the same number of counts in both cases using featureCounts. I was wondering if it is possible to tell featureCounts to exclude duplicates from the counting. Is there an option doing it within Rsubread or do I have to actually remove (as opposed to mark) the duplicate reads?

Thanks

Answer 1

[dbpzdbpz]

The new version of subread (1.4.5) will provide a function to ignore reads or fragments that have the 0x400 flag (the duplicate read flag). This version should be released in days.

[lkmklsmn]

Sometimes in reserach "in days" can turn into weeks, months etc.

Do you know of any official release date for 1.4.5?

[dbpzdbpz]

Sorry that it was delayed for weeks. The 1.4.5 version of subread (inc. featureCounts) was released yesterday on sf.net.

http://sourceforge.net/projects/subread/

Answer 2

[Chris Fields]

It doesn't look as if this is supported, though technically it shouldn't be hard to add (they already parse read bit flags and attributes). Maybe ask the authors? Mailing list is here:

https://groups.google.com/forum/#!forum/subread