Mathematical/Statistical models to predict mRNA half-life

Question

[Shicheng Guo]

Mathematical/Statistical models to predict mRNA half-life

I'm seeking advice on the current best models for predicting human mRNA half-life, including their accuracy and any recent advancements. As part of my research, I aim to understand the dynamics of mRNA degradation and its implications for gene expression regulation. I'm particularly interested in models that offer a balance between accuracy and computational efficiency, as well as those that have been validated with experimental data. If anyone has insights into the state-of-the-art in this field, including relevant publications or databases, I would greatly appreciate your recommendations and any experiences you've had applying these models in your research. Previously some paper mentioned Decay Rates of Human mRNAs: Correlation With Functional Characteristics and Sequence Attributes, however, looks we don't have a systemic way/model to make a great prediction correct?

Thanks,

[LauferVA]

id start by making a list of every known phenomenon that affects rna stability. there will be dozens. youd need accurate readouts of how all of these behave by cell type and by (patho)physiologic state in order to be able to build a model for this.

im not saying its impossible, but it would be a major undertaking ... R01, a U grant, maybe many.

[i.sudbery]

I'm pretty use there are a couple of ML models out predicting RNA stability from sequence. I'm not sure how good they are/generally applicable they are.

https://pubmed.ncbi.nlm.nih.gov/36419176/

Is one example, but it only works on short UTRs from an MPRA assay as far as I can tell.

You might also look at: https://pubmed.ncbi.nlm.nih.gov/36419176/

I'll add that ML models for prediction of RNA stability is something we are also interested in developing.

Answer 1

[Mensur Dlakic]

Have you tried a simple Google search?

https://www.google.com/search?q=predict+mRNA+half-life

Because it has an extra hydroxyl group which increases its reactivity, RNA is inherently more unstable than DNA. Yet RNA stability goes beyond that: it can be increased by circularization, proteins that bind 5' or 3' ends, and by secondary structures in untranslated regions. By the same token, it can be decreased by nucleases or nuclease complexes that specifically target some mRNAs for quick degradation.

I don't think there is yet a systematic model to deal with these problems on a total-RNA scale.