4 ms·
Geometric deep learning of RNA structure (2021)
- lawrenceyan 5y agoIf paywalled - https://daslab.stanford.edu/site_data/pub_pdf/Townshend_Watkins_Eismann_Science_2021.pdf https://daslab.stanford.edu/site_data/pub_pdf/Townshend_Watk...
- dekhn 5y agoIt's safe to say there is a wealth of interesting computational work to do on RNA. It's been very instructive to learn computer science through sequence data analysis, and RNA is one of the most interesting. Most 3D RNA structures can be parsed using fairly simply levels of the chomsky parser hierarchy, but some violate the common rules and require stochastic context free grammars (pseudoknots). Neural methods like this through all that out of the window (at least, afaict this loss function doesn't consider the parse tree of the fold).