questions on duplication rate by RNA-SEQC

Dear support:

I am using the RNA-seqc to evaluate the quality of rna-seq data. The
pari-end data is mapped by Tophat with default parameter.

I am confused by duplication report, which says 100% for "unique rate of
mapped". Actually, there are many reads with "NH:i:3", suggesting three
mapping positions.

Is this a bug in RNA-SEQC, or I understand the "unique rate of mapped" in
the wrong way?

RNA-SEQC version: V1.1.7

Best regards,

RE: questions on duplication rate by RNA-SEQC

Hi,

Sorry for the confusion on this but it's an issue of ambiguous semantics. The uniqueness that the tools refers to is whether there are multiple identical reads at this position. If there are multiple identical reads at this position, the reads are considered 'duplicates' (e.g. PCR or optical duplicates).

You are referring to the uniqueness of the mapping. In other words: is this the only possible mapping position for this read - this is not reflected at all in these metrics. It would be an interesting additional metric, though.

David