AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Protein Structure Tokenization via Geometric Byte Pair Encoding

Michael Sun, Weize Yuan, Gang Liu, Wojciech Matusik, Marinka Zitnik

ICLR 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Generative protein structure design · 6 citations (OpenAlex)

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 glue-angle optimization (differentiable inverse kinematics) (ablation)GeoBPE · Reconstruction, |V|=600 (train)RMSD: 4.39 -> 1.66Improved
add glue-angle optimization (differentiable inverse kinematics) (ablation)GeoBPE · Reconstruction, |V|=600 (train)LDDT: 0.69 -> 0.73Improved
fit vocabulary on task-specific data instead of pretraining data (ablation)GeoBPE · Functional site prediction (average, 8 tasks, 15 splits)AUROC (%): 80.20 -> 80.23No clear effect
fit vocabulary on 1% of pretraining data instead of full (ablation)GeoBPE · Functional site prediction (average, 8 tasks, 15 splits)AUROC (%): 80.20 -> 80.12No clear effect
fit vocabulary on 1% of pretraining data instead of full (ablation)GeoBPE · OOD reconstruction (CAMEO, CASP14)RMSD: 1.77 -> 1.66 (CAMEO, |V|=600)Mixed
increase downstream training data from 20% to 100% (ablation)GeoBPE · Physicochemical property prediction (average, flexibility)Spearman rho (%): 38.24 -> 45.26Improved
increase downstream training data from 20% to 100% (ablation)GeoBPE · Structure property prediction (Homo, average)Macro F1 (%): 35.92 -> 52.21Improved
increase downstream training data from 20% to 100% (ablation)GeoBPE · Functional site prediction (average, 8 tasks, 15 splits)AUROC (%): 79.19 -> 80.20No clear effect
increase M_max from 5000 to 20000 motif occurrences (ablation)GeoBPE · Reconstruction (|V|=6000)RMSD: 1.40 vs 1.39No clear effect
use GeoBPE vocabulary features instead of ESM3 VQ tokensESM3 · Structure property prediction (Homo, average)Macro F1 (%): 36.44 -> 52.21Improved
use GeoBPE instead of ProteinMPNN on flexibility predictionProteinMPNN · Physicochemical property prediction (average, flexibility)Spearman rho (%): 54.41 -> 45.26Hurt
use GeoBPE instead of ProteinMPNN on repeat motif prediction (Fold)ProteinMPNN · Rep (Fold split)AUROC (%): 77.63 -> 56.44Hurt

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