low % of read map to mirase mature miRNAs

Question

[prp291]

I have some small RNA-seq data and I want to predict the miRNA sequence. Genome sequences is available for my organism. After preprocessing like adapter removal, redundant reads removal I choose only 18-24 bp long reads for further analysis. I align this reads with genome and chose only those reads which align with genome for next step. I align reads with RFAM, plant repeats and mRNA sequence of my genome to discard any contamination. Final reads (which are not aligned with RFAM, plant repeats and mRNA) are aligned with mature miRNA sequence obtained from miRBASE database. but bowtie result comes like that

 2562148 (100.00%) were unpaired; of these:
    2560555 (99.94%) aligned 0 times
    264 (0.01%) aligned exactly 1 time
    1329 (0.05%) aligned >1 times
0.06% overall alignment rate

Is it OK to get such a low number of reads? I use this command for bowtie alignment:

bowtie2 -N 0 --un unaligned/unaligned_mirbase_mature.fasta -x mature_mirna/mirbase_mature -f aligned/aligned_to_genome.fasta

Thanks

[h.mon]

What is the protocol for library preparation? Did it involve a poly-A selection?

[prp291]

yes it includes that. Thanks

[David Langenberger]

Do mature microRNAs have a poly-A tail?

[h.mon]

It is not clear to me he is using an appropriate library preparation protocol, but I have little experience with small RNA. And no, I believe mature microRNAs do not have poly-A tails.

[David Langenberger]

Why would you remove redundant reads? That's kinda counterproductive. I highly recommend you to first read publications about small RNA-Seq and microRNAs, before performing any further analysis.

My tip: Think about the protocol and think about the molecules you sequenced, would you expect redundancy?

[manekineko]

He removed redundant reads, thats explains the low result :) he just discard them all and leave one read for each miR

[manekineko]

Be careful with discarding repeats and mRNA hits as some miRNA have transposon or exon origin an you will lose those.

Answer 1

[David Langenberger]

Check this post: miRNA mapping rate is very low.. (less than 0.03%)

[prp291]

thanks for guiding me to that question. Let me try your suggestions first

[David Langenberger]

But please think about your experiment first and do not just follow the suggestions. Probably you have a complete different protocol.... read publications, talk to the wet-lab guys and try to really understand, what they did and read publications! Try an error can be very dangerous in bioinformatics, since you always get results back. If they are good, or f...ed up, you can only decide, when you really know what you are looking at!