Entering edit mode
7.3 years ago
bioinforesearchquestions
▴
370
Hello friends,
I have small RNA sequenced from bovine using Illumina platform. I would like to know what is the procedure for identifying differential miRNA expression profile.
1) First I trimmed the sequenced reads for adapters and kept the reads between 18-35. Is it good enough for downstream analyses?
2) Then aligned the trimmed reads against bovine miRBase database for annotation. I got the output file with the miRNA id, reads with unique matches and reads with multi-matches.
3) Is there any recommended steps and pipeline for identifying differential miRNA expression profile?
Hey, it really comes down to what are you trying to find, if you are only trying to find which miRNAs are DE in your dataset, unlike with mRNA transcriptomes, you just need to grab the raw counts for each miRNA and you are good to go to test for DE with your favorite method (edgeR, DESeq, etc). Perhaps this paper will be helpful to you.
Also I see you tried to emphasize on the first point. What did the sequence length distribution looked like? You should have a peek around 22 nt, but maybe it should be worth it to check the sequences longer than 35nt to see if there is any mistake with the processing you are making.
Hi Biofalconch,
do I need to compare the trimmed miRNA reads against miRBase database or the reference genomes? to generate the raw count?
Sorry for the late response, I have been away for quite a while with no internet access. You could use miRBase to compare your raw data, however you will be missing out on novel microRNAs. You can just map your reads to miRBase and assing the counts on where they map and how well they map (possibly set a threashold of alignment length, mismatches, etc). Also, be careful of microRNAs that are encoded more than once on the genome, they cannot be used to answer certain questions.
Thanks for your response Biofalconch.
I have referred this paper and tool for miRNA expression profiling. I used the mirna_soft tool (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4138881/).
I am working on both bovine and pig.
What do you think about this workflow, from miRBase I extracted bovine and pig related miRNAs from mature.fa (miRBase) file. Using miRNA_soft tool as per the research article I provided in the previous comment, I counted the reads mapped to bovine and pig miRNA sequences extracted from mature.fa. I got an output file with the following details:
miRNA id, miRNA sequence, Unique matches and Multi-matches.
I got similar output for all the samples. Then again using miRNA_soft tool, I compared the counts generated in the previous step to generate differential miRNA expression profiling.
I am a naive person in miRNA sequencing. Do you recommend this workflow?
Do I need to compare against mature miRNA or hairpin miRNA?
You can see this post too What is the acceptable percentage of reads with adapters sequenced in miRNA sequencing?
It seems fine to me, although it is not clear to me what they use to test for differential expression, but having a raw counts file should be enough to work with any differential expression package such as edgeR and DESeq. And I would keep it on the mature miRNAs, since to the best of my knowledge pre-miRNAs do not always correlate with miRNA expression.