diffusion model

Diffusion Latent Representations for Neural Decoding

This study evaluates diffusion latent representations for neural decoding, showing varying performance based on diffusion timesteps.

Diffusion-based Representation Integration for Foundation Models Improves Spatial Transcriptomics Analysis

DRIFT improves spatial transcriptomics analysis by integrating spatial context into foundation models using diffusion on spatial graphs, outperforming specialized methods.

Generative design of antigen-specific T-cell receptor sequences with a conditional diffusion model

This paper introduces TCRDiff, a conditional diffusion model for generating antigen-specific TCR sequences, validated through in vitro assays.

Pep2Mol: 3D Molecule Generation Targeting Protein-Protein Interfaces with Diffusion Models

Pep2Mol generates 3D molecules targeting protein-protein interfaces using diffusion models, achieving state-of-the-art binding affinities for small-molecule inhibitors.

F.A.D.E. (Fully Agentic Drug Engine): A Conversational AI Platform for Drug Discovery

F.A.D.E. is a conversational AI platform that generates drug candidates from natural language queries, outperforming reference ligands in drug-likeness for two targets.

Stimulus and circuit contributions to the information geometry of neural manifolds

The paper develops a differential geometric framework to analyze how neural manifold geometry is shaped by feedforward and recurrent connectivity, showing feedforward connectivity determines representational geometry.

Robust Conditional Diffusion with Noisy Templates for Antibody Sequence-Structure Design

This paper introduces NT-ABDiff, a diffusion model for antibody sequence-structure design that improves CDR recovery and structural accuracy using noisy and unreliable templates.

AFM-Fold: Rapid Reconstruction of Protein Conformations from AFM Images

AFM-Fold reconstructs 3D protein conformations from AFM images using a generative AI framework, outperforming rigid-body fitting and capturing conformational dynamics.

Counterfactual Diffusion Models for Interpretable Explanations of Artificial Intelligence Models in Pathology

MoPaDi generates counterfactual explanations for histopathology images to reveal morphological features driving classifier predictions, improving interpretability in digital pathology.

PertDiffBench: Benchmarking Diffusion Models for Single-Cell Perturbation Response Prediction

PertDiffBench benchmarks diffusion models for predicting transcriptional responses to perturbations in single-cell and bulk RNA-seq data, showing mixed performance across different tasks and settings.

AlphaFold3 for Structure-guided Ligand Discovery

Retrospective and prospective benchmarks show AlphaFold3 co-folding memorizes training poses and achieves half the prospective hit rate of DOCK3 at 6,700-fold higher cost, but excels as a post-docking filter.

An Expert-Informed Synthetic Animal Data Generator: A Physiology-Consistent Generative Framework for High-Fidelity Animal Digital Twins

This paper introduces EICD, a framework that generates physiology-consistent synthetic animal data by embedding mechanistic biological principles, achieving high fidelity and reducing physiological hallucinations.

Chromosome X Dosage Modulates Development of Aneuploidy in Genetically Diverse Mouse Embryonic Stem Cells

The paper identifies that X chromosome dosage influences aneuploidy development in mouse embryonic stem cells, with XO and XY lines showing higher aneuploidy rates than XX lines.

Comparative Evaluation of Deep Generative Models for Capturing Topological Features in Brain Structural Connectivity

This study compares deep generative models for generating brain structural connectivity data, finding that WGAN-GP provides the most stable performance while VAE and DDPM have application-specific strengths.

An Agentic Platform for Drug Repurposing Unified across Molecular, Phenotypic, and Clinical Scales

LinkD is a drug repurposing platform that integrates molecular, phenotypic, and clinical data to prioritize drugs, showing reduced prostate cancer incidence with beta-blockers.

STITCH: Spatial Transcriptomics Imputation via Flow Matching with Internal Learning

STITCH reconstructs high-resolution spatial transcriptomic atlases by imputing missing data in 2D and 3D spatial transcriptomics datasets.

CryoDiff: An uncertainty-aware diffusion model for Cryo-EM map enhancement

CryoDiff enhances cryo-EM maps using a diffusion model with uncertainty estimation, improving resolution and model completeness.

In Silico Generation of Gene Expression profiles usingDiffusion Models

This paper introduces diffusion models for generating synthetic gene expression data and shows they outperform existing methods in predictive performance.

CLOP-DiT: Structured-Metadata-Conditioned Single-Cell Latent Generation via Contrastive Language-Omics Pretraining and Diffusion Transformers

CLOP-DiT generates synthetic single-cell gene expression profiles from structured biological metadata using contrastive alignment and diffusion models, achieving high steering accuracy but limited cross-dataset variability.

Transcriptomics-Conditioned Virtual Tissue Synthesis via Diffusion Transformers

The paper introduces STMDiT, a diffusion transformer that synthesizes H&E histopathology images from transcriptomic data, improving image quality and transcriptomic fidelity over existing methods.

MORPHE: Bridging Image Generation and Spatial Omics for Tissue Synthesis

MORPHE synthesizes biologically faithful tissue architecture from spatial-omics data, enabling reconstruction and extension of tissue context beyond experimental limitations.

Diffusion-based stimulus optimization reveals functional organization across higher visual cortex

The paper validates a diffusion-based method called BrainDiVE for generating stimuli that activate specific brain regions, showing enhanced category selectivity and revealing functional gradients in visual cortex.

From static thresholds to dynamic waves: How biological memory destabilizes malaria transmission Potential

This paper develops a delay-based framework to model malaria transmission dynamics, showing that biological memory from developmental delays can create time-dependent transmission potential and stability fluctuations.

Bayesian-Steered Structure Prediction of Mechanical Biomolecules Using Twisted Diffusion

This paper introduces a Bayesian-steered diffusion method for predicting alternative conformations of mechanical biomolecules, demonstrating its ability to reproduce experimentally validated states of various proteins and DNA.

Simulation-conditioned generative modeling for biologically realistic pattern prediction

This paper introduces a simulation-conditioned generative framework that combines mechanistic models with image generation to create biologically realistic patterns, enabling inference from experimental data.

The Second Brain: Diffusion Models for Realistic Human Microbiome Generation

This paper introduces a diffusion model for generating realistic human microbiome data, achieving high sparsity preservation and competitive ecological distance metrics.

Distance-Restraint-Guided Diffusion Models for Sampling Protein Conformational Changes and Ligand Dissociation Pathways

This paper introduces a distance-restraint-guided diffusion model to sample protein conformational changes and ligand dissociation pathways, achieving uniform coverage and high structural quality.

Image-Conditioned Diffusion for Privacy-Preserving Synthetic Medical Images

This paper evaluates image-to-image diffusion for generating privacy-preserving synthetic medical images, finding it outperforms text-to-image methods in fidelity and clinical signal preservation.

Fast and Ultra-Capable Protein Design: Advancing the Frontier Through Atomistic SE(3)-Equivariance with Genie 3

This paper introduces Genie 3, a diffusion model that improves protein design by combining SE(3)-equivariance with hallucination, achieving state-of-the-art results in binder design and motif scaffolding.

InversePep: Diffusion-Driven Structure-Based Inverse Folding for Functional Peptides

InversePep generates peptides with target structural geometries using a diffusion model, outperforming existing methods in producing geometry-consistent sequences.