diffusion model

A systematic assessment of machine learning for structural variant filtering

This study benchmarks machine learning methods for structural variant filtering in long-read sequencing data, finding that simple models match complex ones in accuracy.

A logarithmic theory of visuomotor stabilization

This paper develops a theoretical framework showing how logarithmic coding enables adaptive visuomotor stabilization in locomotion, validated through analysis of experimental data and extension to intermittent swimming.

Reprogramming BCMA-Targeted CAR-T Cells through γ-Secretase Modulation Blocks Antigen Shedding and Extends CAR-T Longevity

This paper develops SHEDLOCK CAR-T cells that block BCMA shedding, enhancing their cytotoxicity and longevity in multiple myeloma models.

CellFluxV2: An Image Generative Foundation Model for Virtual Cell Modeling

CellFluxV2 is an image-generative model that predicts cell morphology changes under perturbations, achieving higher image fidelity than baselines and enabling batch-effect correction and cell-state interpolation.

PathDiffusion: modeling protein folding pathway using evolution-guided diffusion

PathDiffusion models protein folding pathways using evolution-guided diffusion, accurately reconstructing folding events and demonstrating versatility across multiple applications.

Designing AAV Capsid Protein with viability-guided Diffusion Model

This paper introduces AAVDiffusion, a viability-guided diffusion model for designing viable AAV capsid proteins, showing improved performance in generating functional sequences.

Improved protein binder design using -pairing targeted RFdiffusion

Conditioning RFdiffusion on edge beta-strand pairing yields specific binders to polar targets, with in silico success above 9.2% versus 0.98% for hotspot-only conditioning.

Human Messenger RNA Harbors Widespread Noncoding Splice Isoforms

The paper identifies 15,836 noncoding splice isoforms in human mRNA from 7,298 protein-coding genes, showing tissue-specific expression and association with cancer.

Fast-cWDM Brain MRI: Fast Conditional Wavelet Diffusion Model for Synthesis Brain MRI Modality

This paper introduces a fast conditional wavelet diffusion model for synthesizing missing MRI modalities, achieving competitive performance and high segmentation accuracy.

Predicting the unseen: a diffusion-based debiasing framework for transcriptional response prediction at single-cell resolution

This paper introduces dbDiffusion, a diffusion-based framework for predicting transcriptional responses to unmeasured genetic perturbations using Perturb-seq data, achieving better accuracy than existing methods.

Context-dependent low-dimensional neural dynamics unfold in distinct subspaces, dimensionality, and dynamical strength for natural walking and reaching

The paper compares neural dynamics during walking and reaching, finding that reaching has higher dimensionality and faster decay in dynamical systems than walking.

Diffusion Models vs. DCGANs for Class-Imbalanced Lung Cancer CT Classification: A Comparative Study

This study compares diffusion models and DCGANs for class-imbalanced lung cancer CT classification, finding diffusion models achieve higher accuracy and better recall for benign cases.

Probabilistic Brain MR Image Transformation Using Generative Models

This paper compares probabilistic generative models for MR image transformation, finding NCSN most accurate and DDPM best at detecting out-of-distribution inputs.

OS2CR-Diff: A Self-Refining Diffusion Framework for CD8 Imputation from One-Step Inference to Conditional Representation

OS2CR-Diff improves CD8 biomarker imputation in mIF images by using a self-refining diffusion model with multiple conditional inputs and a feature fusion module.

Influence of natural enemy specificity and functional response on victim coexistence

This study examines how enemy specificity and functional response shapes coexistence of competing victim species using a deterministic model.

Computational enzyme design by catalytic motif scaffolding

This paper introduces Riff-Diff, a machine learning and atomistic modeling method for designing enzymes with high catalytic activity and stereoselectivity, demonstrating its effectiveness for two chemical reactions.

Atom level enzyme active site scaffolding using RFdiffusion2

RFdiffusion2 scaffolds unindexed atom-level catalytic active sites via flow matching without inverse rotamer enumeration, solving all 41 in silico benchmark cases and yielding active de novo enzymes across four reactions.

Computational Design of Metallohydrolases

This paper designs highly active metallohydrolases using a new generative AI method, achieving kcat/KM values up to 53,000 M-1 s-1 without experimental optimization.

All-Atom Protein Sequence Design using Discrete Diffusion Models

This paper introduces a discrete diffusion model for protein sequence design using all-atom representations, showing improved novelty and diversity compared to amino acid-based models.

A dynamical-nonequilibrium molecular dynamics (D-NEMD) alanine scanning approach for identifying allosteric positions in proteins

This paper introduces a D-NEMD method to identify allosteric positions in proteins by analyzing long-range structural responses to alanine substitutions, validated on KPC-2 beta-lactamase.

Can We Extract Physics-like Energies from Generative Protein Diffusion Models?

This paper examines whether diffusion models can extract physics-like energies for protein docking, finding they can recover energy funnels comparable to Rosetta in 9 out of 25 cases.

Keeping SCORE enables interpretable uncertainty-aware classification from diffusion models for genomics

Keeping SCORE transforms diffusion models into probabilistic classifiers for genomics, providing accurate predictions with uncertainty and interpretability across multiple tasks.

BoltzGen: Toward Universal Binder Design

An all-atom diffusion model unifying biomolecular structure prediction and sequence-structure generation designs experimentally validated nanobodies, mini-proteins, and linear or cyclic peptides across 26 diverse targets.

Using Diffusion Transformers to Generate Synthetic Diffusion Scalar Maps for Data Augmentation

This paper generates synthetic brain MRI data using a Diffusion Transformer to augment neuroimaging datasets and improve dementia classification.

Deep generative modeling of temperature-dependent structural ensembles of proteins

aSAM, a latent diffusion model trained on molecular dynamics, generates heavy-atom protein ensembles; its temperature-conditioned variant aSAMt reproduces mdCATH thermal unfolding and extrapolates beyond training temperatures.

Improving Stereochemical Limitations in Protein-Ligand Complex Structure Prediction

This study improves stereochemical accuracy in protein-ligand complex structure prediction by addressing limitations in AlphaFold3 and Boltz-1 through a restraint-guided inference method.

Modeling Phage Therapy

The paper develops a mathematical model to simulate phage therapy success by considering bacterial resistance dynamics and cocktail composition.

SpecLig: Energy-Guided Hierarchical Model for Target-Specific 3D Ligand Design

SpecLig generates target-specific ligands with high affinity and specificity by integrating hierarchical modeling and energy guidance, improving therapeutic potential and reducing off-target risks.

Exploring the Conformational Landscape of Adenylate Kinase and Beyond: A Benchmark of Protein Folding Models

The paper benchmarks nine protein folding models on 20 monomeric proteins, showing Chai-1 excels in recovering conformations and capturing dynamics for Adenylate Kinase.

Comparative Analysis of Diffusion Models for Enhancing Alzheimer's Disease Classification

This study compares two diffusion models for generating synthetic MRI images to improve Alzheimer's disease classification, achieving higher precision and recall than baseline models.