featureCounts overlap settings

[Chi Sutherland]

Hi there,

I am using featureCounts to assign reads to transposable elements, however I was wondering what overlap settings would be recommended for this? I have tried using --fracOverlap 0.5 but have discovered that this limits assignment to very short elements of which there are many in our GTF (unsurprising given that we have paired end 150bp reads and some elements are only less than 50bp) . I also tried --fracOverlapFeature however this limits assignment to larger elements. I am thinking it may be best to use --minOverlap and set it based on read length of datasets, do you think this would be the most unbiased approach or are there any other overlap settings? Is there a minOverlapFeature parameter as this would be ideal! Hope this makes sense, and advice would be much appreciated.

Many thanks,

Chi

[Yang LIAO]

Hi Chi,

The "--minOverlap" option works on both reads and features, namely it gives the minimum number of bases overlapping between the feature and the read, hence there isn't a "--minOverlapFeature" option. 

The best value of it depends on the data in your analysis, say, what are the reads (WGS? Total RNA Sequencing?), how similar the sequences are in the transposable element database, what is the sequencing quality and how did you map the reads, etc.  

Cheers,

Yang