Recently, I have sequenced many libraries on Novaseq. I am aware of the dark cycles causing long polyG reads due to Novaseq's two color technology. However, I am also seeing pretty consistent overrepresentation of CA repeats and GT repeats representing between 0.2-0.6% of my total reads. It is also important to mention that these libraries were made using oligodT priming, so I shouldn't really be seeing a lot of short tandem repeats represented in the libraries. Is this something that other people are seeing, and is it an artifact of the technology or is it a potential contamination in the data?
The quality of my reads is good, and I'm not super worried about whether these need to be trimmed out or not because STAR usually does a great job mapping regardless of trimming, I'm just more curious about why I am seeing this occur regularly.
Hey,
may I ask about the quality of these specific repeats? Do you have a drop in quality here?
Regards,
André
Andre,
Thanks for the response. In answer to your question, I ran the following to extract one of the overrepresented sequences (put here for my future reference and for anyone else who might need this):
zgrep -B1 -A2 "^CACACACACACA" lib1_R1_001.fastq.gz | grep -v "^--$" > overrep.fastq
This extracted the CA repeats into its own Fastq file. I then ran FastQC to determine whether quality scores are lower for these reads than for the average reads in the original fastq file. While certainly not universally so, there is a clear drop in quality score for the CA repeat reads (and the same for the GT repeats), particularly at the end of the reads. I've attached all of the pertinent FastQC files here.
Does this point to something to you? Should I chalk this up as an artifact of the sequencing, or something more problematic?
Best,
Jim
- Badges
- Science topic
- Similar topics
- Correction