Thanks for the response! The adaptor content ranges between ~10-30% on the tail end of the reads in the samples. Heres a representative sample that we had:

This is why I began incorporating cutadapt into my workflow for processing these samples. I'm fairly confident that it's not primer dimers as all of the samples were ran on bioanalyzer and I'm working under the impression that a pretty clear peak would show up if we were getting primer dimers. Also, even if there were primer dimers I'm assuming that they would be removed in the cutadapt step. If it helps heres the command I'm running for that:

cutadapt \
-a AGATCGGAAGAGCACACGTCTGAACTCCAGTCA \
-A AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT \
-j 10 -o AC05trimmed.R1.fastq.gz -p AC05trimmed.R2.fastq.gz \
AC05_1.fastq.gz AC05_2.fastq.gz \
-m 15 

and the output for one of the samples:

Total read pairs processed: 117,228,555
Read 1 with adapter: 43,501,641 (37.1%)
Read 2 with adapter: 42,090,827 (35.9%)
Pairs written (passing filters):  117,228,555 (100.0%)

Total basepairs processed: 35,403,023,610 bp
Read 1: 17,701,511,805 bp
Read 2: 17,701,511,805 bp
Total written (filtered): 33,299,771,975 bp (94.1%)
Read 1: 16,629,715,004 bp
Read 2: 16,670,056,971 bp

In terms of ribosomal contamination I tried to get this info but I'm not sure if the way I did it works. Basically I tried generating an rRNA bed file from the pre-existing gtf file I had by filtering for entries that contained 'gbkey "rRNA"' (I'm using the GRCm38 reference). Then I used bedtools intersect to find all of the reads in my bam file that overlapped with my bed file:

bedtools intersect -a GRCm38Aligned.sortedByCoord.out.bam -b B6.rRNA.GRCm38.bed -wa > GRCm38.rRNA.Aligned.sortedByCoord.out.bam

and then I ran idxstats from samtools to get a summary of the output. From this I only got around 2.8 M reads. If the way I got to this number is wrong let me know.

Assuming that it's not primer dimers or ribosomal contamination, does that just mean the samples are duds? Also does 130 M read pairs just mean that we sequenced at a very high depth? Sorry I still very new to all this, so I'm trying to get a better idea of how everything works. Thank you again for the help!!

130M reads is fairly typical for a bulk total RNA prep really. You pick up a ton more transcripts without the polyA selection, and if you want to look at alternative splicing or identify transcripts de novo, the additional reads really help.

However, I would be slightly wary of diagnosing the problem as PCR duplication. With only 8 rounds of PCR, the maximum number of decendants of an original molecule is 256, yet in that sample you have

100K copies (some of these won't be true copies, as this is single end duplication), but even so, this is probably too high.

Good analysis. Although he seems to just be guessing that it was 8.

Anyway, RNA-seq is rife with duplicates. To get rid of some of the artificial ones, you can use BBTools:

clumpify.sh in=reads.fq out=deduped.fq dedupe optical ddist=50 passes=6

The specific distance that you use for optical duplicate removal is platform-specific. 50 is fine for NovaSeqX, NextSeq, and HiSeq 2000, 1T, and 2500. Consult the shellscript documentation for other platforms, it's highly variable...

That's generally a safe operation to reduce Illumina-specific artifactual duplicates. You can always get rid of all duplicates with "dedupe" leaving out "optical" but that makes no sense for quantitative RNA-seq experiments.

Doing any kind of PCR amplification on quantitative RNA-seq experiments will bias your results, so it should be avoided if possible, but if you have to do it, you have to do it. You can't remove the artifacts with digital deduplication, though, without compromising your results.

Dear @jnechacov thank you for joining our community.

We note that all of the posts you have added in the 12 hours since you joined mention a particular product. Note that while there is no community ban on recommending products which you are associated with, associations must be excplicitly disclosed for posts not to be marked as spam.

Secondly, the use of generative AI in composing posts is again, not outside the rules of the community, but posts using generative AI must be specifically useful to the question or risk being deleted as off-topic. Your answer has a structure common to answers from many generative AI tools, such as chatGPT. Your answer here is superficially relevant, so I'll leave it up for now. But be aware that obviously GPT generated answers are not likely to engender confidence in your product recommendations.