Biomedical Text Mining and Ontologies

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

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.

This paper introduces a guided AI framework for automating data harmonization to the OMOP Common Data Model, achieving high agreement with human experts while enabling scalable and efficient data integration.

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

megaMine extracts gene-cancer-drug evidence from biomedical literature using a rule-based framework, achieving high accuracy in distinguishing efficacy from non-efficacy evidence.

The paper introduces BioChirp, a system for reliable retrieval of structured biomedical data, outperforming existing agentic NL2SQL systems in accuracy and consistency.

PhysioMap connects physiological traits to causal relationships and predicts responses in metabolic diseases, showing agreement with references after excluding discordant cases.

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

MAOMAO integrates and harmonizes peptide toxicity data from 54 sources to create a FAIR resource with standardized terminology and benchmark partitions for toxicity research.

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.

BiomarkerKB integrates and standardizes biomarker data from multiple sources to enable structured exploration and discovery of biomarker-disease relationships.

HARMONY harmonizes metadata across metabolomics repositories, improving cross-repository search and study comparison by resolving metadata inconsistencies using ontologies.

PrimeKG-Plus updates the PrimeKG knowledge graph with new data and rare-disease information, improving drug-disease connectivity and temporal relevance.

CellExLink improves end-to-end cell-type recognition and Cell Ontology normalization in biomedical text, outperforming existing tools across diverse corpora.

GeneKnow synthesizes biological evidence from literature with auditability, achieving higher claim support and citation fidelity than existing AI systems.

The paper introduces ECO, an AI-native ontology that unifies evolutionary and cell biology in latent space, demonstrating its effectiveness in simulating protein evolution, cell cycles, and latent-space alignment.

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.

Interaction-finder automates the discovery of biological entity associations from literature, producing a ranked list with quoted provenance, and outperforms existing methods in recall.

VirProtRAG improves viral protein function annotation by integrating literature retrieval and evidence-based generation, introducing 32.53% new annotations and supporting sequence-inferred functions.

RD-OMICS integrates and structures rare disease omics data from GEO into a knowledge graph to support research and drug repurposing.

The paper presents an AI workflow for mapping Key Events to genes in Adverse Outcome Pathways, improving accuracy and reducing manual effort through semantic retrieval, LLM refinement, and expert curation.

The paper uses Vaccine Ontology to model immune biomarkers, profiles, and responses, enhancing classification and analysis of vaccine-induced immunity.

PhenoXtract extracts phenotypes from clinical descriptions using a hybrid approach of Large Language Models and Knowledge Graph embeddings, achieving high recall and precision.

trAIt uses large language models to retrieve species traits from scientific literature, achieving 47.1% accuracy in a case study, though expert curation remains more accurate.

The paper introduces RECoDe, a relation extraction dataset for diet, disease, and biomarker associations, and benchmarks NLP models on it, achieving high classification performance.

The paper develops deep learning models for microbial named entity recognition and normalization to accelerate literature review and meta-analysis of microbiome studies.

MetaHarmonizer improves metadata harmonization and addresses LLM contamination in biomedical data, outperforming existing methods on benchmark tasks.

This paper creates a structured atlas of research logic in Plant Cell by converting articles into Research Process Graphs using LLMs, revealing patterns in research methods and findings.

This paper introduces KM-GPT-DCH, a method using LLMs and Bayesian analysis to evaluate and compare biomedical hypotheses by analyzing scientific literature.

Ignet 2.0 and Vignet provide ontology-driven platforms for gene interaction discovery, vaccine network analysis, and AI-assisted knowledge retrieval from biomedical literature.