mRNAs targeted by list of miRNAs

Question

[Megatron]

Hi,

I have three lists of microRNAs that change significantly (each consisting of 20-30 microRNAs)

For each list, I want to find what the top messenger RNAs (mRNAs) predicted targets.

Something similar to this is on microRNA.org (where I choose mouse, input my mRNAs, then click on "view mRNAs targeted by ALL selected miRNAs"

However, for some of my lists that have too many miRNAs, microrna.org finds no genes that overlap. Instead, I would like a way to find the best genes that overlap the most, or somehow score the highest among a list of chosen miRNAs.

I have also explored DIANA miRPath, where a list of entered miRNAs are enriched into pathways. This, however, only shows me the genes that are involved in the KEGG pathways of interest.

I am interested in rounding down a large list of miRNAs into a small number of top mRNA targets which I can analyze using qPCR.

Any tips?

Many thanks

Megatron

[Lorena Pantano]

Hi,

This is quite challenging due to the large false positive you get when getting targets without having more information. If you do only this, probably your 'top genes' will be biased to long 3UTR genes and most study miRNAs.

I will try to get all targets for each miRNA, count how many miRNAs target a gene, divide bu length of the utr, and use the one with higher score. This still is dangerous, because probably there are genes that are not even expressed on your tissue of interest. If you can get the potential genes, better. I don't know if you have a phenotype change or not, but you can focus as well the search on those pathways that you think there is something.

As far as I know, there is no way to call the best genes with only miRNA data (confidently enough to select for PCR and have a high TP rate). If you had mRNA expression as well, then you can do many other things.

[Megatron]

"If you had mRNA expression as well, then you can do many other things." Did you mean miRNA expression data or mRNA expression data? (mRNA expression data is exactly what I would seek if I would know where to look - which is the purpose of my next step being discussed here)

Cheers