Cleaning paired-end Nextera prepped reads

[rhinoceros]

These reads were sequenced (2x100) with Hiseq 2500 rapid mode. In the first pictures, the reads as they were when I got them:

R1




R2





And then I ran trim_galore (0.3.7). I used the reverse complement of the 3'-end of the Nextera® transposase sequence as an adapter sequence:

trim_galore -q 20 --fastqc -a CTGTCTCTTATA --stringency 1 --paired --retain_unpaired --length 50 $1 $2

And this is what the reads look like now:

R1:




R2:




I assume all the bias that is left in the 5' and 3'-ends is due to the cutting preferences of the Nextera transposase. Is this assumption correct, i.e. are my reads clean?

[Brian Bushnell]

That's correct. You should be fine, though it does look like you gained a weird bias on your last 1bp. However, fastqc makes it hard to tell if that's 1bp or several. I assume that you are overtrimming the reads by removing even 1p of sequence that matches the adapter, or something like that.

[rhinoceros]

That's correct. You should be fine, though it does look like you gained a weird bias on your last 1bp. However, fastqc makes it hard to tell if that's 1bp or several. I assume that you are overtrimming the reads by removing even 1p of sequence that matches the adapter, or something like that.

Thanks for the reply. About the 3'-end, I'm indeed most likely overtrimming. In trim_galore:

--stringency <INT> Overlap with adapter sequence required to trim a sequence. Defaults to a very stringent setting of 1, i.e. even a single bp of overlapping sequence will be trimmed off from the 3' end of any read.

Anyway, relatively little information is lost because of this, so there's no harm. I was more concerned about the 5'-end, since I never saw what trimmed Nextera were supposed to look like. I also thought it was a little bit weird how the Nextera transposase sequence started appearing from the middle of the reads onwards but I guess that's normal..

[nucacidhunter]

I also thought it was a little bit weird how the Nextera transposase sequence started appearing from the middle of the reads onwards but I guess that's normal

Your plots indicates that around 15% of library fragments had shorter insert than number of sequencing cycles. If the library size cut off was a bit larger, Nextera adapters would not have been sequenced which would provide more useful data.

[rhinoceros]

Your plots indicates that around 15% of library fragments had shorter insert than number of sequencing cycles. If the library size cut off was a bit larger, Nextera adapters would not have been sequenced which would provide more useful data.

I have to admit that I'm a little bit ignorant on the technical details of the sequencing part. Is library size cut off related to fragment size distribution? Also, what is the relation of fragment sizes and the number of sequencing cycles? Also, the DNA comes from a metagenomic sample which was very small (6.8ng), so overall I'm quite happy about the output..

[yueluo]

If you have a smaller fragment size than read length, you will get adapter sequence on the 3' -end of your reads. I think you can try running Brians' bbmerge to merge reads by overlapping.