Haipeng Gong

Decision benchmarking reveals when protein foundation models add experimental value

This paper benchmarks decision strategies for protein engineering, showing when models and measurements improve decisions based on accumulated evidence.

qcGEM: a graph-based molecular representation with quantum chemistry awareness

qcGEM introduces a quantum-chemistry-aware graph-based molecular representation that improves performance and interpretability in molecule-related tasks.

From single-sequence structure prediction to protein fitness landscape through a composable, epistasis-aware mutation atlas

The paper introduces Cerebra-Epistasis, a framework that improves multi-mutant fitness prediction by modeling epistatic effects and integrating structure-aware sequence analysis.

Constructing the ensemble of representative structures for a protein via neural-surrogate-guided MSA recombination

This paper introduces ProCEDiS, a framework that generates representative protein conformations using neural surrogates and AlphaFold2, validated on four protein systems.

qcMol: a large-scale dataset of 1.2 million molecules with high-quality quantum chemical annotations for molecular representation learning

Accelerating Virtual Directed Evolution of Proteins via Reinforcement Learning

This study introduces RelaVDEP, a virtual directed evolution pipeline using reinforcement learning to accelerate protein mutation discovery, showing improved catalytic activity in PETase through novel mutations.

Improving diffusion-based protein backbone generation with global-geometry-aware latent encoding

This paper introduces TopoDiff, a diffusion-based framework for protein backbone generation that improves global geometry control and fold diversity through a structure encoder and latent sampler.

GDFold2: a fast and parallelizable protein folding environment with freely defined objective functions

GDFold2 is a protein folding environment that enables fast, parallelizable folding with customizable objective functions and a quality assessment model.

Title: Cell-autonomous and non-cell-autonomous effects of TRIM72 on ALS disease progression

This study identifies TRIM72's protective role in ALS through both cell-autonomous and non-cell-autonomous mechanisms, showing its therapeutic potential in animal models and human patients.

Improving the prediction of protein stability changes upon mutations by geometric learning and a pre-training strategy

The paper introduces three geometric learning models for predicting protein stability changes upon mutations, achieving improved performance over existing methods.

SPIRED-Fitness: an end-to-end framework for the prediction of protein structure and fitness from single sequence

SPIRED-Fitness predicts protein structure and fitness from single sequences using an end-to-end framework combining SPIRED with downstream neural networks.

Cerebra: a computationally efficient framework for accurate protein structure prediction

Cerebra improves computational efficiency in protein structure prediction while maintaining accuracy, achieving comparable performance to optimized models with fewer resources.

Identification of residue pairing in interacting β-strands from a predicted residue contact map

This paper introduces RDb2C, a method to identify β-strand residue pairings from predicted contact maps, improving β-protein structure prediction.