question about overlap between JACUSA2 and REDItools

[gianfilippo]

Hi,

this is more of a question

I ran JACUSA2 and REDItools on the same sample (RNA and DNA). I find little overlap between the calls. I see about 25% of the JACUSA calls shared with REDItools.

Is this expected or does it seem unexpected ?

Thanks

[piechottam]

No, this is not expected.

Did you check the default parameters for minimal BASQ, minimal mapping quality of each Tool? How do you calculate the overlap? -> try unstranded overlap. What type of sites are missing in JACUSA2 output? What is their coverage?

[gianfilippo]

Hi, thanks. Read/base Quality are set to be the same, 20. I created a basic bed fie from REDItools2 output and intersected it with JACUSA2 bed output file (bedtools intersect -a $REDIfile -b $JACUSAfile) I used REDIportal to annotated and select known events calls REDItools2 I also used REDIportal (after converting it to a 0-based bed file) as input to JACUSA2 For instance, I get the following N of calls REDItools2 314490 JACUSA2 157450 intersection 37438

I understand the filters may not be fully equivalent, but I was expecting ore calls in the intersection. That said, I feel I am making a mistake somewhere. What do you think ? Thanks

[piechottam]

Ty. That overlap is strangely low. Do you mind sending me the command lines for each tool?

[gianfilippo]

Hi, I am very sorry, I reviewed the code and reran the contrasts and intersection is actually very good. Best