In silico identification of miRNAs

Question

[z.zangishei]

Hi everybody

An easy issue that made me confused these days is why ESTs are used for in silico identification of miRNAs or why other types, for instance transcriptome seq data, may not be suitable for?

Answers would be highly appreciated.

[Buffo]

I´m not an expert, but; If you are looking for target prediction you have to use RNA-seq data for better results, or at least count just miRNAs mapping to CDS. But, for some species there is no RNA-seq data, and then you can use ESTs (even when this in not the most appropriate) If you are looking for the location where the miRNA was synthesized you have to use the whole genome sequence. At least in mouse the majority came from non-conding regions.

[z.zangishei]

Thanks for your consideration Buffo

Actually I have different goal of what you mentioned. I want to predict miRNAs of my interest organism computationally. As I could understand from references, mostly they have used ESTs and/or gss for homology searches. Now I can not understand why ESTs are used? and how about other data like transcriptome sequencing data?

[Buffo]

ESTs are used just because they are from mRNAS and also because sometimes it is everything you have, in silico miRNA prediction is a complicated issue, technically and theoretically. It is based on finding hairpin sequences along CDS, and then predict DROSHA and DICER proccessing.. etc etc. If I where you, I will start making miRNA-seq, it is more expensive than in silico prediction but you will get more significant information.

[z.zangishei]

As a matter of fact in silico predictions are complicated approaches, but are used widely specially as initial actions. Regarding miRNA-Seq, presently there is not the possibility to do that.

As I can understand you are believe that every data related to mRNA can be used for the miRNA prediction, am I right? However, I think they have used ESTs even though when transcriptome sequencing data were available, like tomato and etc.

[Buffo]

Using ESTs data when RNA-seq is available does not have any sense, ESTs are from random mRNA sites.