Physics-based AI
for RNA engineering

Physics-grounded models for predicting, designing,
and optimizing RNA structures across research workflows.

RNAccess by Emergente

RNAccess is Emergente's on-demand platform for RNA structure prediction.

1
Submit your RNA sequence
2
Generate physics-grounded 3D structural predictions
3
Explore your structures

Our Products

Live now

RNAccess

A webserver for fast RNA structure prediction. Move from sequence to structural insight — no setup required.

High AccuracyEasy to UseIntuitive
RNAccessLive
The RNAccess 3D viewer: a predicted RNA fold coloured 5′→3′, with ΔG −27.43 kcal/mol.
Coming soon

RNAissance API

Integrate the RNAnneal engine directly into your pipelines.

ReliableScalableSecure

Powered by RNAnneal

Best-in-Class RNA Structure Prediction Engine

RNAnneal is our engine that fuses physics simulation with generative AI to fold any RNA from scratch. It ranked #1 in 3 of 4 structural accuracy metrics and posted the highest interaction fidelity across the benchmark.

Benchmarked on 24 challenging RNA targets
CASP15 CASP16 RNA3DB
#1 · Best in class
0.96
Interaction Network Fidelity (INF)
↑ Higher is better

How faithfully the predicted base-pair and base-stacking network matches the experimental structure.

INF, higher is better: RNAnneal 0.96 (best), Boltz-2 0.83, AlphaFold3 0.87, Chai-1 0.84.
Best in class
RMSD
↓ Lower is better
RMSD, lower is better: RNAnneal 9.9 Å (best), Boltz-2 10.4, AlphaFold3 11.7, Chai-1 13.8.

The average distance between atoms in a model and the reference structure after alignment.

Competitive
eRMSD
↓ Lower is better
eRMSD, lower is better: RNAnneal 1.41, Boltz-2 1.33, AlphaFold3 1.35, Chai-1 1.52.

The average distance between nucleobase centers in a rotation- and translation-invariant representation.

Best in class
emDI
↓ Lower is better
emDI, lower is better: RNAnneal 1.58 (best), Boltz-2 1.70, AlphaFold3 1.60, Chai-1 1.98.

A deviation index that amplifies differences in 3D structure when the interaction network is broken: the eRMSD divided by the INF.