REVEAL: RNA Ensemble Validation and Exploration with AI-guided Landscapes

REVEAL captures the reversible ES2↔GS register shift in the engineered HIV-1 TAR mutant, recovering both basins from 6.8 s aggregate unbiased sampling across 34 seeds.

• Why it matters: Excited RNA states are sparsely populated and rare, so unbiased molecular dynamics from one structure often cannot reach the needed transitions or reproduce correct state populations.
• What they did: The workflow combines generative secondary-structure sampling, secondary-structure conditioned trRosettaRNA2 tertiary prediction, unbiased molecular dynamics, and Markov state modeling, then evaluates ensemble quality with PWCS.
• The result: Simulations seeded from the NMR ES2 ensemble remain trapped in ES2, while PWCS ranks conformers to identify the dominant state even when low-dimensional representations fail.

AI summary of the abstract · may contain errors · not medical advice · read the paper.