AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

NC-Bench and NCfold: A Benchmark and Closed-Loop Framework for RNA Non-Canonical Base-Pair Prediction

Heqin Zhu, Ruifeng Li, Ao Chang, Mingqian Li, Hongyang Chen, Peng Xiong, S Kevin Zhou

ICLR 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Generative protein structure design

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RNA foundation model embeddings (REF) to the base model (ablation)NCfold-base · Pair-EdgeF1: 0.251 -> 0.365Improved
add RNA foundation model embeddings (REF) to the base model (ablation)NCfold-base · OrientationF1: 0.419 -> 0.486Improved
add BPfold base-pair motif energy prior (ablation)NCfold-base · Pair-EdgeF1: 0.251 -> 0.335Improved
use RFM embeddings instead of BPfold energy prior (ablation)NCfold-BPE · OrientationMCC: 0.326 -> 0.312Mixed
increase transformer layers from 2 to 6 (ablation)NCfold · Pair-EdgeMCC: 0.131 -> 0.219Improved
increase transformer layers from 6 to 10 (ablation)NCfold · Pair-EdgeF1: 0.336 -> 0.317Hurt
increase training batch size from 4 to 64 (ablation)NCfold · Pair-EdgeMCC: 0.207 -> 0.061Hurt
increase training batch size from 4 to 32 (ablation)NCfold · OrientationMCC: 0.358 -> 0.207Hurt
select top-2 RFM embeddings instead of top-1 (ablation)NCfold · Pair-EdgeF1: 0.341 -> 0.365Improved
select top-3 instead of top-2 RFM embeddings (ablation)NCfold · Pair-EdgeF1: 0.365 -> 0.336Hurt
fine-tune NCfold head instead of frozen RFM with linear layerfrozen RFMs · Pair-EdgeF1: 0.218 -> 0.365Improved
use NCfold instead of canonical-pair predictors in zero-shot NC pairingBPfold · NC pairing status (zero-shot)recall: 0.021 -> 0.431Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy