FastQC Report

[hamcan]

I ran a HiSeq on environmental samples and the purpose of the run was to blast my sequences against the NCBI-nr database to see what species my reads match to. I am not doing a denovo assembly or genome assembly.

My FastQc report passes in all aspects except: per base sequence content, per sequence GC content and kmer content.
Should I be worried? How much should I rely on a fastqc report?

Thank you in advance!

[mastal]

I would worry most about the per base sequence content, depending on what it looks like, why it didn't pass.

[hamcan]

I would worry most about the per base sequence content, depending on what it looks like, why it didn't pass.

Below are the pictures of the FastQC failed reports:

[mastal]

The per base sequence content looks OK, you might need to trim the last few bases at the 3' ends of the reads. The kmer plot looks like there might be adapters at the 3' ends of the reads too.

[Zapages]

I would trim up the first 19 bps at the 5' end (which probably are the adapters) and trim the last 50 bps at the 3' end.

Also I would suggest increasing the kmer count to k 10 in FastQC to get a better idea of things for the 3' end for how much to trim.

All the best with your project.

-Zapages

[Brian Bushnell]

I would trim up the first 19 bps at the 5' end (which probably are the adapters) and trim the last 50 bps at the 3' end.

I think not; Nextera libraries normally look like that at the beginning due to shearing bias, but the bases are correct. The 3' end looks like adapter sequence, though, and should be adapter-trimmed.

[GenoMax]

I would trim up the first 19 bps at the 5' end (which probably are the adapters) and trim the last 50 bps at the 3' end.

Also I would suggest increasing the kmer count to k 10 in FastQC to get a better idea of things for the 3' end for how much to trim.

All the best with your project.

-Zapages

No trimming necessary. Refer to this post by Dr. Simon Andrews, author of FastQC.

[Zapages]

No trimming necessary. Refer to this post by Dr. Simon Andrews, author of FastQC.

Very interesting development and something that I always thought about this too when I was working on my data sets.

Since the biased composition is created by the selection of sequencing fragments and not by base call errors the only effect of trimming would be to change from having a library which starts over biased positions, to having a library which starts slightly downstream of biased positions.

Thank you for sharing.

I did a lot of RNA-Seq analysis last year and earlier this year. This news was not known at that time... When I free time, I definitely will go back and check some of my old results and see if there is any improvement in my differential expression results.

Whilst the warnings generated by this problem reflect a real issue it’s not something which can be fixed, and doesn’t seem to have any serious consequences for downstream analysis. Ironically if you are producing RNA-Seq libraries it would make for better QC if you were to focus on libraries which didn’t have this artefact in them, as they would be the ones which were truly suspicious.

I guess, we should go with more expensive PCR-free approaches: https://konradpaszkiewicz.wordpress....biased-genome/

Thoughts?

Would you recommend this approach for older generated data that used TruSeq Library Prep kits or had 5' that were really messy? As I think back, I remember dealing with some pretty messy RNA-Seq that had to be cleaned up from Illumina HiSeq 2500 machines. I will give my old results another look when I am free.

[hamcan]

I think not; Nextera libraries normally look like that at the beginning due to shearing bias, but the bases are correct. The 3' end looks like adapter sequence, though, and should be adapter-trimmed.

Hey, it was adapter trimmed at the 3' end! So i'm not sure what is going on..suggestions?

[hamcan]

No trimming necessary. Refer to this post by Dr. Simon Andrews, author of FastQC.

This would explain the 5' end, but how about the 3' end?

[Brian Bushnell]

You may have used the wrong adapter sequences, or simply had incomplete trimming. I suggest starting with the raw reads and performing adapter-trimming as in the post I linked, then looking at the results.