ExperimentBank || miRNA-gene interaction network

Experiment Bank - The site for free exchange, review, and discussion of novel scientific experiments

Front Page ? Everything ? News ? Wish list ? Help ? FAQ

?miRNA-gene interaction network
By RH ZHANG, Section Biology Posted on Fri Apr 30th, 2010 at 01:31:52 AM PST

In recent years, non-coding RNAs are attaining more and more interest, especially micro-RNAs. These oligo strands of RNA have gradually come to be realized as a whole new system of gene expression regulation. With advanced sequencing technologies, there is already about a thousand sequences of human miRNA annotated in miRBase, one of the most frequently consulted archives for miRNA research. However, the sequences have yet to give rise to full understanding, as the significance of most miRNAs are still unknown. With few target-validated miRNAs, researchers still rely heavily on prediction algorithms based on sequence chemistry or inferred from co-expression profiles, such as the often used miRanda, miRBase targets (now Microcosm), TargetScan, etc. However, they are highly prone to false-positives and false-negatives (Barbato et al, 2009). Recently, Costanzo et al. (2010) have used a synthetic genetic array (SGA) methodology to map gene-gene interactions in yeast. It incorporates systematically induced double gene mutants in S. cerevisiae followed by assessments of the organism's viability in comparison to two separate single mutants of the same two genes. This provided a deduction of the interaction between two genes, which may be positive or negative. Following this approach, I propose that we take advantage of the technique of synthetic lethality to do the same for a miRNA-gene network, where the second mutation would be in a miRNA. However, it would be more practical to introduce exogenous miRNA rather than a mutation at miRNA sites, as miRNAs may have a diversity of production origins. And since miRNAs are mostly negative-acting, the assessment of combined effects of the miRNA and gene might be able to give rise to a network of gene-miRNA interactions. This would be highly useful as it provides a functional view of miRNAs, as opposed to the pure sequence-target relationship we have been focusing on.

< Pharmacogenomics in Africa (2 comments) | The evolution of Pangaea (6 comments) >

miRNA-gene interaction network | 6 comments (2 topical, 4 editorial)

Experimental design none (#1)
by Kendric on Tue Apr 27th, 2010 at 02:01:37 PM PST
(User Info)

Interesting idea, but I have a few concerns.

Like you mentioned, the original "double synthetic lethality" technique was demonstrated in yeast. Are you proposing to utilise the technique for miRNA-gene interactions also in yeast? If so, do you know how common miRNA regulation occurs in yeast? (Perhaps I wasn't looking in the right place, but I couldn't find any entries for s.cerevisae)

I may have misunderstood your intention, but I thought the goal of this experiment was to find the genes reguated/transcripts targeted by miRNA. I'm not sure if this experiment can elucidate these interactions.

This is how I understand the experiment to work...
(1)When only geneX is mutated, we may observe some decrease in viability.
(2)When only miRNA is overexpressed (which represses its targeted mRNA(s), we may observe some decrease in viability.
(3) When both miRNA is overexpressed and geneX is mutated,
we may observe greater decrease in viability than expected if (1) and (2) are independent (not interacting). This tells us that the single mRNA (or multiple different mRNA) acted on by the interacts with the mutated gene. However, we still do not know what transcript(s) are being acted on by the miRNA..

Regarding finding miRNA interactions, a paper has recently been published using a technique called Argonaute HITS-CLIP.

Chi, S. W., Zang, J. B., Mele, A. & Darnell, R. B. Argonaute HITS-CLIP decodes microRNA-mRNA interaction maps. Nature 460, 479-486 (2009)

there's no miRNA is yeast. none (#2)
by RH ZHANG on Wed Apr 28th, 2010 at 09:40:59 AM PST
(User Info)

You are absolutely right about the organism, this wouldn't be possible in yeast. In fact, due to the scale, I think it would only be practical to do this in vitro, on a mammalian cell line. But my intention is certainly not to find gene targets of miRNA, as that is the work of target-validation of predicted targets. That kind of interaction can be done by the various new IP techniques, like the one you mentioned. However, due to the fact that each single miRNA has the potential to affect hundreds of genes, the kind of work I'm talking about is more suited to determining the effects of overall functional interactions between the miRNA in question under the condition of a mutated well-known gene. This would provide additive information to existing studies of single miRNA effects and has the potential to elucidate the miRNA's role at the cell level. The goal is not to study it at the molecular level, since the mutated gene transcript is not necessarily acted on by the miRNA.

[ Parent ]

RE: there's no miRNA is yeast none (#3)
by timothy auyeung on Wed Apr 28th, 2010 at 10:23:33 AM PST
(User Info)

I really like this proposal. But I am having some difficulties understanding your last comment. If what we are interested in is the global effects of a given miRNA on the cell and that (as you have pointed out) the mutated transcript may no longer be affected by the miRNA (eg the miRNA binding site on the transcript is mutated), why do we need to mutate the transcript in the first place? Can we not simply overexpress just the miRNA? I understand that you wish to investigate the effects of miRNA under the condition of the mutated gene. However, even if we observe a phenotype, we may not know whether the phenotype is caused by 1) miRNA alone, 2) mutated gene or 3) a combination of both since (again) we do not know whether the interaction between the miRNA and the transcript is preserved after the mutation.

Nevertheless, I think this is a great proposal :)

[ Parent ]

Interaction between Gene A and Gene Group none (#4)
by RH ZHANG on Wed Apr 28th, 2010 at 12:51:40 PM PST
(User Info)

Thank you for bringing this up. Now you've mentioned it, I feel that there are things I should have added to my initial proposal. It is certainly true that if we wanted to investigate the global effects of the miRNA we would just overexpress it. Now with the approach I'm proposing, we are actually asking the question: What are the effects of a miRNA expression under the condition of mutated Gene A. Or the other way around: What are the effects of a mutated Gene A under the condition of a miRNA silencing a group of genes.

So now I thought more about it, this is more like taking Costanzo et al.'s (2010) approach of Gene A vs Gene B to a slightly different level: now we are talking about the interaction between Gene A and the outcome of a miRNA expression; in other words, the interaction between Gene A and a group of co-regulated genes.

Say, for example, this miRNA is predicted or validated to downregulate 50 genes. The effect of this experiment would be the outcome when mutated Gene A interacts with these 50 genes. This is excellent because we already know these 50 genes are regulated by the same thing so we can treat it as a Gene Group. Then we compare this to the phenotypes when there is 1) miRNA alone and 2) mutated gene alone.

What would be more interesting is when we compare this to the results of each individual gene-gene interactions. We can take the measure of change from Costanzo et al.'s (2010) Gene A interaction with each of the 50 regulated genes. Our expectation would be that the Gene A - Gene Group interaction be the sum of all the individual 50 Gene A - gene interactions. Any deviation from that expectation would tell us about the combined effect of these 50 interactions, which is in essence the effect of a miRNA - Gene A interaction.

[ Parent ]

[new] Being done in C. elegans (none / 0) (#5)
by Earthworm Jim on Wed Apr 28th, 2010 at 12:52:45 PM PST
(User Info)

RNAi is consistently use in C. elegans (best model ever!) and even in our lab, synthetic lethal interactions are being worked out through RNAi VIA feeding.

[new] Nice idea! (none / 0) (#6)
by CourteneyLai on Fri Apr 30th, 2010 at 01:31:43 AM PST
(User Info)

Just wanted to comment that I like the idea of trying to improve prediction abilities as, as you mentioned, the current in silico methods are notoriously inaccurate or unable to be experimentally validated!

miRNA-gene interaction network | 6 comments (2 topical, 4 editorial)

?+?Menu

? create account
? faq
? search

?+?Login

Make a new account

Username:

Password:

?+?Related Links

? More on genomics
? Also by RH ZHANG

All trademarks and copyrights on this page are owned by their respective companies. Comments are owned by the Poster. The Rest ? 1999 The Management

create account | faq | search