Quantify miRNAs from whole transcriptome library (single end, TruSeq, 100bp)

Question

[w.denhollander]

I'm trying to wrap my head around the quantification of miRNAs from a (ribosome depleted) whole transcriptome library. Most of the approach I find here (and elsewhere) focus on specifically sequenced small RNAs.

Could anyone point me in the right direction? Preferably using Rsubread, but other approaches (Bowtie2 for example) are also fine.

EDIT: I should add that the alignment and quantification of gene transcripts seemed to work out fine using Rsubread, but I understand that as miRNAs are substantially smaller than my 100bp reads, I have to do some magic. I tried indexing and aligning against mature.fa from miRBase, but I end up with 0 aligned reads.

Answer 1

[i.sudbery]

Just in case you (or anyone else) comes back to this. Usually we would think that we can't qualitfy miRNAs from whole transcriptome sequencing because the size selection step in library prep process will specifically remove any miRNAs from the sample. Thus, while whole transcriptome libraries are called "whole transcriptome", they generally only capture that part of the transcriptome bigger than about 100bp. There is no magic that can recover molecules that phyiscally arn't in the sample.

Answer 2

[GenoMax]

I have to do some magic

There is no magic. Find out what kit was used for the library prep. Normally there is a specific adapter involved that is directly ligated to miRNA' before library prep. This adapter is kit specific and will need to be trimmed before aligning the data (using ungapped alignments so use bowtie v.1.x). You will want to find out what kit was used.

[w.denhollander]

Thanks for the reply!

All the information I have is the following, but I guess that wouldn't be sufficient?

The libraries were prepared using out whole-transcriptome CORALL v2 kit with ribodepletion (RiboCop). The read length was SR100.

TruSeq adapters are used in the libraries: Read 1: AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC

[GenoMax]

Confirm this is the kit before proceeding.

https://www.lexogen.com/wp-content/uploads/2023/07/171UG394V0111_CORALL-RNA-Seq-V2-with-UDIs_2023-07-12.pdf on page 36 has instructions for data analysis.

This appears to be a total-RNA kit so you will have things other then miRNA (unless something special was done).

Answer 3

[markus.glass]

Usually, miRNAs are sequenced using specific smallRNA-library preparation methods (I think most, if not all of these methods include a size selection for small RNAs). I guess, the main problem with quantifying miRNAs from a totalRNA-seq is that miRNA abundance might be biased due to an excess of other RNA species getting most of the sequencing reads. Thus, I guess except for some very abundant miRNAs, e.g. miR-21 or let-7 (if you've sequenced human samples), results might not be very acurate.

Nevertheless, now, that you've got the (3'?) adapter sequences from the library prep kit, you can trim those off your reads using tools like Cutadapt or Trimmomatic, then align the sequenced reads to your reference genome/transcriptome using e.g. Bowtie and then quantify using e.g., Rsubread/FeatureCounts and the miRBase gff3 as annotation base.

Answer 4

[jnechacov]

Quantifying miRNAs from a ribosome-depleted whole transcriptome library can be challenging because traditional library prep methods aren't optimized for small RNAs. Here are some suggestions to improve alignment and quantification:

Read Length and Trimming:

Since miRNAs are ~20-24 nt, aligning 100 bp reads directly to mature miRNA sequences can lead to poor results. Use tools like Cutadapt or Trimmomatic to trim adapter sequences and shorten your reads to match miRNA lengths before alignment.

Aligning with Bowtie2:

Bowtie2 is well-suited for short reads. When aligning to the mature.fa file from miRBase, try setting parameters for very short, exact matches:

bowtie2 -x miRBase_index -U trimmed_reads.fastq -L 15 -N 0 --n-ceil L,0,0.15 -k 1 -S miRNA_alignment.sam

The -L 15 parameter adjusts the seed length, and -N 0 allows no mismatches in the seed.

Using Rsubread:

Rsubread isn’t typically optimized for small RNA alignment, but if you prefer it, ensure your reads are trimmed to ~20-24 nt before aligning. Also, consider setting a smaller fragment length and allowing more mismatches in your alignment parameters.

rRNA Depletion Optimization:

Residual ribosomal RNA can impact the detection of low-abundance miRNAs. To improve your rRNA depletion step in future experiments, consider using Zymo Research’s PureRec Duplex-Specific Nuclease (DSN). PureRec DSN efficiently removes rRNA and highly abundant transcripts, enhancing the detection of miRNAs and other low-expression targets, resulting in cleaner, more informative libraries.

Quantification Tools:

After alignment, use tools like miRDeep2, mirge2.0, or featureCounts (if using Rsubread) to quantify miRNA expression accurately.

By optimizing trimming, alignment, and rRNA depletion with PureRec DSN, you’ll likely improve your miRNA detection from whole transcriptome libraries. Hope this helps point you in the right direction!

[i.sudbery]

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 answer from many generative AI tools, such as chatGPT. Your answer here is superficially relevant, so I'll leave it up for now, but it doesn't really get the the heart of the users problem, which is getting miRNA reads out of a total RNA seq library.