M. Baek

Essential and virulence-related protein interactions of pathogens revealed through deep learning

This study uses deep learning to predict protein-protein interactions in bacterial pathogens, identifying new complexes involving essential genes and virulence factors, with half of tested predictions validated.

Generalized Biomolecular Modeling and Design with RoseTTAFold All-Atom

This paper introduces RFAA and RFdiffusionAA, models that can predict and design complex biomolecular assemblies including proteins, nucleic acids, small molecules, and covalent modifications.

Protein Ensemble Generation through VariationalAutoencoder Latent Space Sampling

The paper generates high-quality protein conformations using a variational autoencoder, achieving better sampling accuracy than MD simulations or AlphaFold2 for K-Ras.

Efficient and accurate prediction of protein structure using RoseTTAFold2

This paper introduces RoseTTAFold2, a protein structure prediction method that achieves AlphaFold2-level accuracy without using its key features, demonstrating the potential for diverse model architectures.

Identification of hyperosmotic stress-responsive genes in Chinese hamster ovary cells via genome-wide virus-free CRISPR/Cas9 screening

The paper identifies hyperosmotic stress-responsive genes in CHO cells using CRISPR/Cas9 screening and validates their role in stress resistance and protein production.

Protein-protein interactions in the Mla lipid transport system probed by computational structure prediction and deep mutational scanning

The paper maps protein-protein interactions in the Mla lipid transport system using deep mutational scanning and AlphaFold2 structure predictions, revealing overlapping binding sites on MlaC.

Accurate Mutation Effect Prediction using RoseTTAFold

This study evaluates RoseTTAFold's ability to predict mutation effects without additional training, achieving state-of-the-art accuracy across diverse protein families.

Hallucinating protein assemblies

Deep network-based protein design generates symmetric homo-oligomers with high structural accuracy, demonstrating the potential for creating novel nanomachines and biomaterials.

Improving de novo Protein Binder Design with Deep Learning

This paper improves de novo protein binder design by using deep learning models to assess design success rates and computational efficiency.

The {varphi}PA3 Phage Nucleus is Enclosed by a Self-Assembling, 2D Crystalline Lattice

The paper shows that the Phage Nucleus of {phi}PA3 forms a self-assembling, 2D crystalline lattice with adaptable inter-tetramer contacts and functional channels.

Little skate genome exposes the gene regulatory mechanisms underlying the evolution of vertebrate locomotion

The paper generates a high-quality little skate genome and analyzes motor neuron gene regulation across species to identify conserved and divergent mechanisms in vertebrate locomotion evolution.

Deep learning methods for designing proteins scaffolding functional sites

This paper introduces two deep learning methods for designing proteins with specific functional sites, showing they can generate diverse proteins with desired properties.

Structures of core eukaryotic protein complexes

The paper uses coevolution analysis and deep learning to model core eukaryotic protein complexes in yeast, identifying 104 new assemblies and structurally characterizing 608 previously unknown complexes.

Toward More General Embeddings for Protein Design: Harnessing Joint Representations of Sequence and Structure

This paper introduces a data-efficient method for protein embeddings by jointly training on sequence and structure, improving mutation effect prediction through enriched representations.

Structure of the phosphoinositide 3-kinase p110γ-p101 complex reveals molecular mechanism of GPCR activation

The paper reports the cryo-EM structure of the p110γ-p101 complex, revealing how GPCR activation is mediated through interactions with Gβγ and identifying potential therapeutic targets.

Class I Histone Deacetylases (HDAC1-3) are Histone Lysine Delactylases

This study identifies HDAC1-3 and SIRT1-3 as enzymes that remove histone lactylation marks, contrasting with the known role of acetylation.

Learning a force field from small-molecule crystal lattice predictions enables consistent sub-Angstrom protein-ligand docking

The paper improves protein-ligand docking by optimizing force field parameters using small-molecule crystal structures, achieving better binding structure prediction.

Mechanism-based inhibitors of SIRT2: structure-activity relationship, X-ray structures, target engagement, regulation of alpha-tubulin acetylation and inhibition of breast cancer cell migration

The paper develops potent, selective SIRT2 inhibitors with features suitable for cell culture and therapeutic development.