This is a test version of Biostars. For the public version, visit https://www.biostars.org.
bcl2fastq results in Poly-N in R1

Hi, we're working on a scRNA-Seq samples, where R1 has 16bases with the cellular and molecular barcodes, while R2 is 150bases long and contains the genomic sequence.

This data set appears to be very problematic, as it shows many problems. We think that we have sequence into the adapter, as running fastqc shows an over-representation of PolyA stretches and a drop of quality as one can see in the first attachment. R2 polyA tail

For that reason we would probably hard trim the samples (maybe after qc-trimming)before the quality drop. But what we don't understand is why our R1 shows only stretches of N.

PolyN-Stretches

Does anyone has an explanation for this kind behavior? Has anyone seen something like this before? the bcl2fastq did not show any errors at all and the fastq is just a long list of read with ployN stretches.

thanks Assa

bcl2fastq paired-end rna-seq scrna-seq single-cell
• 2,509 views • link
updated by GenoMax GenoMax • written by Assa Yeroslaviz Assa Yeroslaviz

Do you have access to the library QC? (pre-sequencing)

• 0 views • link
written by Asaf Asaf

1 answer

You need to use option --mask-short-adapter-reads 0 with bcl2fastq to prevent the short first read from getting masked. Reads less than 25 bp are normally masked with N's.

• 0 views • link
written by GenoMax GenoMax

this might be the correct solution. thanks

• 0 views • link
written by Assa Yeroslaviz Assa Yeroslaviz

The reason for this is that in 10X fragments there is a polyT stretch right after the barcode/UMI part of the read. Once the machine hits this stretch base quality will rapidly drop and will not recover. This does not matter though since BC/UMI comes before this stretch. Therefore you can ignore this fastqc result for R1. Still, this is the reason why you should not mix "normal" NGS libraries with 10x samples because the normal samples would suffer from that quality drop. For 10X the R2 is what actually reads the RNA/cDNA and for this the fastqc result is relevant.

• 0 views • link
written by ATpoint ATpoint

This is true, but this is not 10X samples and the polyA stretches are in R2, not R1.

• 0 views • link
written by Assa Yeroslaviz Assa Yeroslaviz

Log in to answer this question.

More posts like this