HiSeq run results

[sajoshi]

Attached is a picture of the HiSeq 2500 run I recently did. The data looks really good. But the user says "I haven’t mapped to the reference genome, but it does not look good. See the sequence duplication levels of FastQC report. Only 4% seq left after removing duplications. It supposed to be over 40%"

What does this mean? And is he analyzing it the right way? This is RNA seq data. Please explain. Not sure If he is aligning properly with the right software.

[fanli]

Agree, run metrics look good. What kind of library prep was it? Sounds like the input complexity was low to start, or perhaps overamplified. Inefficient rRNA removal would also be a possibility.

Duplication rate can be assessed directly on the sequencing reads without mapping.

[GenoMax]

Attached is a picture of the HiSeq 2500 run I recently did. The data looks really good. But the user says "I haven’t mapped to the reference genome, but it does not look good. See the sequence duplication levels of FastQC report. Only 4% seq left after removing duplications. It supposed to be over 40%"

What does this mean? And is he analyzing it the right way? This is RNA seq data. Please explain. Not sure If he is aligning properly with the right software.

Please ask the user to read Dr. Simon Andrew's blog post on duplication metrics in FastQC here.

You user would need to provide additional information for us to be of more specific help. I am not sure why the user is trying to de-duplicate RNAseq data since they are throwing away important count information and may in fact mess the results up. That said, I agree with @fanli that this data may not have been ribo-depleted and may be full of rRNA (which is > 90+% of cellular RNA). Problems are likely to be with the user samples than the sequencing.