Non-coding RNA prediction using genomic data

Question

[nancydong20]

Hello everyone!

I am trying to analyze non-coding RNAs in a newly assembled genome. I had tried Infernal (as reported in published studies), but our collaborator recommended to predict different types of non-coding RNAs (ex. miRNA, tRNA, rRNA) separately using more dedicated tools? For example, predict tRNAs using tRNAscan-SE (which I did and completed), predict miRNAs and rRNAs using BLAST with the respective databases.

I have tried to use BLAST with the miRBase, but I am getting very short alignments and I'm not sure how meaningful it is.

What is the best way to analyze non-coding RNAs in genomic sequence (which I know isn't going to be very accurate anyways)?

Thank you very much!

[biofalconch]

As far as miRNAs go, you could use miRDeep2 in order to get microRNAs, but be advised, you will need small RNA-seq data to get it up and running.

Answer 1

[cschu181]

You could scan your genome one more time with Infernal http://eddylab.org/infernal/...

Edit: failed to see that OP already used this and now trying to get out of an embarrassing situation with humour. :P

[GenoMax]

Looks like OP has already done that.

[cschu181]

Blargh. I didn't see that.

[nancydong20]

No worries! I wonder why the collaborator recommended that I scan for them separately. Is it going to be more sensitive if I do it separately?