Low quality sequencing from too short R1

[fxjonas]

Hi,

Our lab started using the mcSCRBseq protocol for RNA library preparation (https://www.protocols.io/view/mcscrb...otocol-p9kdr4w), which recommends a 16/8/0/50 read settings in the Illumina machine 16= BC+ UMI, 50= cDNA. Interestingly, the read quality with these settings became significantly worse than when we used our previous libraries with 50/8/8/16. (% >Q30 dropped from 85-95% to 65-75% and increased phrasing 0.1 -> 0.3). After contacting Illumina about this problem, they mentioned that a too short R1 could cause problems when determining good clusters and result in problems later on. This sounds very conclusive, but unfortunately, we are not able to adjust the length of R1 and actually test his prediction. Therefore I was wondering, if anyone has experience with short Read1 sequencing and if someone knows a way to improve the quality of the reads.
Thanks for your help,
Felix

[GenoMax]

You just have to live with the quality of the reads you get. Illumina requires 25 cycles in first read to accurately calibrate Q scores. Your type of sequencing is common with "Drop-seq" like methods where the first read is short as in your case.

I am not sure if the read structure in your case can allow you to sequence longer on read 1. If it can you could sequence 25 cycles and then drop the 7 bases.

[ajanjic]

Hi Felix,

Thanks for giving mcSCRB-seq a try. I think this may be instrument dependent. We have access to a HiSeq 1500 and we have not had issues with low Q30 values, as for this particular instrument 16 bp seems to be sufficient for establishing clusters.

It may be impractical for you to increase your read length on R1 though, because then you will read into the polyA tail which would impact your quality.


Best,
Aleks Janjic