How to avoid microRNA-like off target effects using computational methods

Question

[Li Hsing]

Hi all,

I am trying to build a Rat shRNA database, but I don't known how to avoid microRNA-like off target effects using computational methods?

Thanks

Answer 1

[Jean-Karim Heriche]

If you're concerned about miRNA-like effects then you need to consider the seed region of your interfering RNA. You may find some inspiration from this paper : A bioinformatics method identifies prominent off-targeted transcripts in RNAi screens.

Or you could apply miRNA prediction methods to your shRNAs. For starting points see this post: Question: Best algorithms for mammalian microRNA target prediction

Answer 2

[CraigM]

You can create a blast database of rat mRNA sequences and blast the sequence against it and parse the results to check that it has no off target effects in genes other than your intended target.

Use blast settings such as this to find small matches such a string of 16 nucleotides which may allow miRNA binding.

blastn -db rat_mrna -word_size 16 -evalue 1000 -query ratshrna.fa > ratshrna.bln

I implemented something like this in a siRNA finder script which can be found at:

https://github.com/cmonger/CHO_siRNA_Finder/blob/master/syfy.pl

[Jean-Karim Heriche]

miRNAs work through imperfect matching of a 6-8 bases seed region to 3'UTRs of mRNAs. Even in the general case, using a word size of 16 would risk missing relevant matches. For example, see this paper: Many commonly used siRNAs risk off-target activity