De novo RNA Backbone Generation using DDPM

Recently with the rise of large language models the world of computational biology has exploded (especially in terms of proteins). Despite this, insufficient tools exist for engineering RNA in silico.

Comment

Inspiration

Inspired by developments in de novo protein design (specifically RosettaFold Diffusion)

What it does

Generates canonical RNA backbones in PDB format. Trained off the RCSB protein databank, is designed to mimic characteristic qualities of native RNA models solved in vitro with large variability.