fastqc acceptable file

Question

[Morris_Chair]

Hi guys, We hired a service for sequencing RNA of several samples for different set of experiments. I performed the qc before and after the trimming and clipping of files but one to me is not good at all because there are lots of errors

Please find attached the link for the download of the fstqc

https://www.dropbox.com/s/0uehk6qrew2rfha/fastqc.html?dl=0

when I do the alignement I have 22% of read alignment against the 80% of the rest of the samples.

here is the answer of the company

For 22% of the reads aligning (vs. 80% for the other samples), it would seem like 1/4 of the reads align. I'm not quite sure how this affects your analysis, but based on the samples passing all of our internal QC, the read output was nearly 4x the minimum (~30 million reads versus 8 million reads), and the quality of the reads was very good, we wouldn't expect there to be a problem with this sample.

It's true that we paid for 8M of reads and we got more, some samples are 20M other 30M (which I m not sure if this will effect the result as well)

Should I trust what the service says?

I did a HeatMap but I mean...how can this be ok if it's completely different

https://ibb.co/41XNjy1

thank you

Answer 1

[lakhujanivijay]

First, about the fastqc results. Looking at the quality boxplots, I can say that it is "ridiculously good" data however having worked for a couple of NGS projects, I never seen such a plot where the yellow box plots just collapse which means there is little or no variation in the phred score at particular read location. Generally the plot looks something like this for good quality data. Hence, I am a little sceptical.

Then, let's talk about "more" number of reads. Now, having more data is good but that does "not" always mean "good" downstream results.