large language model

PubMind: Literature-Based Genetic Variant Extraction and Functional Annotation Using Large Language Models

PubMind extracts and annotates genetic variant functionality from biomedical literature using LLMs, creating a database with over 1.3 million variants.

An AI System for Autonomous Algorithm Evolution in Drug Development

DrugEvolve autonomously evolves algorithms across drug development stages, improving performance on benchmark tests and showing generalizability across data modalities.

Automating scientific annotations for open transcriptomic profiles via multi-stage agents

GEOMeta automates metadata curation for GEO transcriptomic data, generating standardized annotations for 600,000 human bulk RNA-seq samples and benchmarking LLMs for annotation quality.

Spatial second-order features predict glioma malignant transformation

This study identifies second-order spatial features, such as ligand-receptor interactions, as predictive of glioma malignant transformation using integrated spatial DNA and RNA profiling.

ACCREDIT: A Quality-Aware Agentic Engine for Cell-resolved Cross-modal Image Registration with Dynamic Iterative Tuning

ACCREDIT improves cross-modal image registration by detecting failures and automatically recovering alignments, outperforming existing methods in accuracy and robustness.

LinkLlama: Enabling Large Language Model for Chemically Reasonable Linker Design

LinkLlama improves linker design by combining natural language prompts with 3D spatial awareness, achieving higher chemical validity than existing methods.

A cross-modal generative model for incomplete and degradedprostate MRI with multicentre clinical validation

This paper develops MSCNet, a cross-modal generative model for reconstructing and restoring prostate MRI sequences, showing improved image quality and diagnostic performance in multicentre validation.

Automated generation of a gene perturbation transcriptomic atlas using large language models

The paper develops an automated pipeline using large language models to generate a gene perturbation transcriptomic atlas from public data, achieving high precision and recall in identifying perturbation groups.

Human-supervised Agentic AI for Hypothesis Generation and Experimental Assistance in Drug Repurposing

A human-supervised multi-agent LLM system covered drug repurposing from knowledge graph building to assay analysis in three case studies; retrospective COVID-19 screen AUC-ROC 0.986 without predefined thresholds.

BioMaster: Multi-agent System for Automated Bioinformatics Analysis Workflow

BioMaster automates complex bioinformatics workflows using a multi-agent system, improving accuracy, efficiency, and scalability across tasks like RNA-seq and Hi-C processing.

Queryome: Orchestrating Retrieval, Reasoning, and Synthesis across Biomedical Literature

Queryome is a deep research system that improves biomedical literature synthesis through iterative reasoning and retrieval, achieving higher accuracy than existing methods.

Decoding Prokaryotic Whole Genomes with a Product-Contextualized Large Language Model

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

Schema-Grounded Multitask Instruction Fine-tuning for Joint Biomedical Named Entity Recognition and Relation Extraction in Pharmacovigilance

This paper fine-tunes large language models for joint biomedical named entity recognition and relation extraction, achieving improved performance across multiple tasks and datasets.

Scientific computing in the age of agentic AI: an exploratory field report

This paper explores the use of LLM-based coding agents in scientific computing, presenting case studies in the life sciences and discussing their potential and challenges.

GeneInsight: Condensing Gene Set Knowledge via Language Models

GeneInsight synthesizes gene set annotations using topic modeling and language models to provide interpretable summaries that reveal biologically significant insights overlooked by traditional enrichment analyses.

DrugLM: A Unified Framework to Enhance Drug-Target Interaction Predictions by Incorporating Textual Embeddings via Language Models

DrugLM enhances drug-target interaction predictions by integrating textual embeddings from large language models into DTI models, showing strong performance without fine-tuning.

Detailed curation of biological samples and experimental designs for genomics using LLM-supported agentic workflows

The paper describes a software tool using LLMs to automate data curation for genomics, achieving performance near human curators with lower cost and higher speed.

PG-MLD: Physics-Guided Molecular Representation Learning via Dynamic 3D Trajectory Distillation

PG-MLD transfers dynamic 3D physical knowledge to SMILES-based molecular models, improving property prediction without requiring 3D conformer generation.

Benchmarking biochemical networks generated by large language models

This study benchmarks large language models' ability to generate biochemical networks, finding moderate accuracy in signaling and metabolic networks but variable prediction performance.

LLM-powered Functional Gene Set Summarization with genesetGPT

genesetGPT is an LLM-based framework that improves gene set annotation by using curated biological context and iterative prompt design to summarize heterogeneous gene sets.

BioPathfinder: Evidence-guided multi-agent platform enables hypothesis discovery for CAR-T engineering

A provenance-tracked multi-agent system curated 57 CAR-T patient scRNA-seq studies, generated reviewable hypotheses, and prioritized KLRC1/NKG2A, whose blockade improved CD8+ CAR-T antitumour activity in a Raji xenograft model.

Integrative AI-Enabled Virtual Cell Modeling Reveals a Clinically Relevant Latent Effector State of Human CD8 T Cells Undetectable by Conventional Analyses

The paper uses AI to model CD8 T-cell responses during ICI therapy, identifying a latent effector state that conventional methods miss.

MechAInistic: An LLM-guided Multi-Agent System for Reasoning over Genome-Scale Constraint-Based Metabolic Models

MechAInistic uses a multi-agent system guided by LLMs to generate auditable hypotheses from genome-scale metabolic models, identifying potential drug targets and metabolic rewiring in disease states.

IOBRpy enables agentic multi-omics decoding of anti-tumor immunity

Python toolkit chains RNA-seq quantification, six-method tumour microenvironment (TME) profiling, SpecHLA typing and TRUST4 repertoire analysis with AI agents; TME outputs match R implementations (Spearman 0.986 to ~1.0).

Evo 2's Perception of Single Nucleotide Substitutions in the Genes of Two Plant Model Organisms

Evo 2 7B assigns higher perplexity to plant gene alleles carrying randomly simulated single-nucleotide substitutions than to those observed in vivo, in Arabidopsis and rice; significant but small effect.

Using large language models for enhancing accessibility for Monte Carlo photon transport simulations and beyond

This paper develops an LLM interface for scientific simulations, achieving high accuracy and reproducibility in converting natural language descriptions to structured inputs for photon transport simulations.

BioReason-Pro: Advancing Protein Function Prediction with Multimodal Biological Reasoning

A reasoning LLM combining ESM3 embeddings with GO-GPT term predictions produces protein function annotations that experts rated tie-or-better than UniProt in 79% of cases (SFT model).

PeptiVerse: A Unified Platform for Therapeutic Peptide Property Prediction

PeptiVerse provides a unified platform for predicting therapeutic peptide properties using both amino acid sequences and SMILES-based representations, supporting early-stage development and generative design.

Biological Continued Pretraining Reshapes the Capability Profile of a Foundation Model Without Catastrophic Forgetting

This paper re-analyzes a foundation model's checkpoints to show biological continued pretraining improves general knowledge, code generation, and reasoning without catastrophic forgetting, but subsequent fine-tuning reverses these gains.

VFB-MCP: Natural-Language Access to Drosophila Neuroscience Grounded by an Expert-Curated Ontology-Led Knowledgebase

Model Context Protocol tools over Virtual Fly Brain's graph let Claude answer 25/30 Drosophila neuroscience tasks precisely, vs 14/30 with web search, 2/30 bare.