protein design

This study explores how oligonucleotide design affects G4 ligand-conjugated oligonucleotide binding and stabilization of G-quadruplex structures.

Pro4S predicts protein solubility by integrating sequence, structure, and surface data, improving accuracy and reducing non-expressed proteins in de novo design.

The paper introduces faster variants of the Caprola tool for calcium-dependent fluorescent labeling, enabling more sensitive and efficient recording of neuronal activity in zebrafish.

This paper develops yeast-selected SpCas9 variants with re-specified PAM recognition to enable allele-specific editing, demonstrating improved specificity and catalytic efficiency compared to PAM-broadened variants.

This paper introduces MP4, a transformer-based model that translates natural language into functional de novo proteins, achieving high performance in sequence realism, fold quality, and function alignment.

The paper develops a fluorescent reporter, SeARCh, to quantify selenocysteine incorporation and reports a hybrid biosynthetic pathway achieving 90% Sec incorporation.

The paper improves the binding efficiency and interaction stability of thermophilic cohesin-dockerin pairs on bacterial cell surfaces using computational and experimental methods.

This paper develops a computational framework to prioritize de novo protein binders for small molecules based on ligand physicochemical properties and evaluates its effectiveness across different ligand types.

The paper introduces GenSyntax, a large language model for prokaryotic genomes that captures functional context, achieving competitive performance in genome-scale tasks.

This study compares different fine-tuning strategies and classical machine learning approaches for protein function prediction using benchmark datasets.

The paper introduces SAKe, a designer protein that forms stable, pH-dependent two-dimensional assemblies for functional protein-based materials.

This paper introduces InterAb and InterAb-Opt, computational frameworks that predict and optimize broadly neutralizing antibodies using all-atom modeling and pre-trained language models, showing significant improvements in binding affinity and neutralization activity.

This paper designs a de novo protein fold that binds doxorubicin with high affinity through aromatic π-stacking, revealing a new 5-helix globular fold.

This paper introduces ALSEBO, a Bayesian optimization method for protein design that efficiently finds optimal sequences with few evaluations by leveraging coevolutionary statistics.

This study identifies a CHCH-helix interface between CHCHD10 and TDP-43CTD that modulates their assembly equilibria and aggregation behavior.

This paper introduces EvoMOBO, a multi-objective Bayesian optimization framework for protein engineering, showing improved performance in enzyme tasks compared to wild-type variants.

This study investigates how folding mRNA into a six-helix bundle affects translation, finding that only the start of the coding sequence needs to be accessible, and that modifications like 5-methoxyuridine and N1-methylpseudouridine influence translation levels.

This study uses zero-shot models to predict activity-enhancing mutations in fluorinase, leading to improved catalytic efficiency and stability in engineered variants.

MAXWELL calibrates protein language models to predict mutation-induced stability changes, outperforming existing methods and successfully designing stabilizing mutations in DhaA dehalogenase.

This paper uses a machine learning model to generate new anti-PD-1 antibodies with improved potency, validated through molecular dynamics simulations and experimental testing.

This paper retrained ProteinMPNN on halophilic proteomes to design salt-tolerant enzymes, showing HaloMPNN shifts protein properties toward halophilic traits.

This paper introduces seed-guided diffusion-based design to improve structural diversity and target contact in antitoxin protein binders, achieving functional binders with enhanced selectivity.

This paper demonstrates the use of bacteriophage P22 virus-like particles as nanoscale protein scaffolds for plant synthetic biology, showing their ability to assemble, encapsulate, and display proteins in Nicotiana benthamiana.

This paper introduces CLSS, a contrastive learning model that co-embeds protein sequences and structures, revealing evolutionary relationships and structural transitions.

This study combines machine learning and directed evolution to optimize the monooxygenase HpaBC for improved L-Dopa production activity.

This paper introduces Boltz2ESI, a framework predicting enzyme-substrate interactions using a foundation model, outperforming existing methods by capturing active-site plasticity and evolutionary context.

This study identifies ENAH as an interactor of the HTT1a PRD and shows that EVH1 domain binding modulates HTT1a aggregation and fibril remodeling.

TriFlow, discrete flow-matching sequence designer with global three-track structure attention, raises in silico binder success over ProteinMPNN in RFdiffusion and BindCraft-style pipelines on 500 targets and flags functional sites.

A parallelized BindCraft pipeline generated 306,146 in silico binder candidates against 8,296 human proteins, each with sequence, predicted complex and confidence metrics, freely accessible online.