Hello,

I'm analyzing a bulk paired-end 150bp RNA-seq data with approx 35M paired reads per sample. dupRadar is showing relatively low duplication rate for genes with high expression level (see the scatter plot - https://freeimage.host/i/dhW3Cu).

Could it be an indication of a problem or simply the use of longer reads decreases the probability to identify a read as duplicated?

Thank you very much for your answer!

Hi Oleg,

your scatter plot looks normal for a paired end experiment. Good paired-end experiments show a mild slope because read pairs with same start in read 1 may still have different end in read 2. This reduces the probability of 2 reads coming from different fragments and having exactly the same start and end (definition of a duplicate).

cheers,

Sergi