Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Implementing trust in non-small cell lung cancer diagnosis with a conformalized uncertainty-aware AI framework.

Nature biomedical engineering·2026

Same author

Generating synthetic multi-national longitudinal cohorts for clinically grounded HIV research.

Nature communications·2026

Same author

Seasonal Variation and Genetic Evaluation of Needle Catechin Content in Half-Sib Families of <i>Pinus taeda</i>.

Plants (Basel, Switzerland)·2026

Same author

A real-world feasibility evaluation of LLM-based clinical prediction: emergency department return visit admission across two academic medical centers.

Research square·2026

Same author

Evaluating Large Language Models for Translating Multimodal Phenotype Documentations into Executable EHR Phenotyping Algorithms.

Research square·2026

Same author

Evaluating Large Language Models for Translating Multimodal Phenotype Documentations into Executable EHR Phenotyping Algorithms.

medRxiv : the preprint server for health sciences·2026

Same journal

CoAff-DTI: Fine-grained drug-target interaction prediction using pre-trained language models and affinity-guided mechanisms.

Journal of biomedical informatics·2026

Same journal

Evaluation of temporal preservation in synthetic longitudinal patient data.

Journal of biomedical informatics·2026

Same journal

ARKE: An ontology-driven framework for automated mapping of local radiology procedure terms to the LOINC-RadLex playbook using large language model.

Journal of biomedical informatics·2026

Same journal

A validation-driven training controller for cross-lingual biomedical NER via reinforcement learning-based adaptive loss weighting.

Journal of biomedical informatics·2026

Same journal

ASP-HR: An Adaptive Spatial Perception and Hierarchical Reasoning mechanism for document-level biomedical relation extraction.

Journal of biomedical informatics·2026

Same journal

Beyond Accuracy: Safety-Centered guidelines for the evaluation of LLM-based therapy recommendation systems for chronic multimorbidity patients.

Journal of biomedical informatics·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Apr 26, 2026

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

Limestone: high-throughput candidate phenotype generation via tensor factorization.

Joyce C Ho¹, Joydeep Ghosh¹, Steve R Steinhubl²

¹Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX 78712, United States.

Journal of Biomedical Informatics

|July 20, 2014

Summary

This summary is machine-generated.

Limestone, a new method using tensor factorization, identifies patient phenotypes from electronic health records with minimal supervision. This approach efficiently derives clinically meaningful phenotype candidates, improving data-driven research.

Keywords:

Dimensionality reduction EHR phenotyping Nonnegative tensor factorization

More Related Videos

High Throughput Yeast Strain Phenotyping with Droplet-Based RNA Sequencing

High Throughput Yeast Strain Phenotyping with Droplet-Based RNA Sequencing

Published on: May 21, 2020

Production and Optimization of LTE, a Leishmania tarentolae Derived Cell-Free Protein Expression System for Recombinant Protein Production

Production and Optimization of LTE, a Leishmania tarentolae Derived Cell-Free Protein Expression System for Recombinant Protein Production

Published on: November 8, 2024

Related Experiment Videos

Last Updated: Apr 26, 2026

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

High Throughput Yeast Strain Phenotyping with Droplet-Based RNA Sequencing

High Throughput Yeast Strain Phenotyping with Droplet-Based RNA Sequencing

Published on: May 21, 2020

Production and Optimization of LTE, a Leishmania tarentolae Derived Cell-Free Protein Expression System for Recombinant Protein Production

Production and Optimization of LTE, a Leishmania tarentolae Derived Cell-Free Protein Expression System for Recombinant Protein Production

Published on: November 8, 2024

Area of Science:

Biomedical Informatics
Data Science
Clinical Research

Background:

Electronic health records (EHRs) offer vast data for clinical research but often lack direct mapping to medical concepts.
Current EHR phenotyping methods are frequently labor-intensive, require expert supervision, and are organization-specific.
Existing approaches are often disease-centric, limiting broader application in clinical research.

Purpose of the Study:

To introduce Limestone, a novel nonnegative tensor factorization method for unsupervised EHR phenotyping.
To demonstrate Limestone's ability to derive phenotype candidates by analyzing patient diagnoses and medication interactions.
To showcase the simultaneous identification of multiple phenotypes from EHR data with minimal human intervention.

Main Methods:

Utilized nonnegative tensor factorization to model interactions between diagnoses and medications within EHR data.
Represented data source interactions using tensors, a generalization of matrices.
Applied Limestone to a large cohort of longitudinal patient records from the Geisinger Health System.

Main Results:

Limestone successfully identified phenotype candidates, revealing patient clusters based on diagnoses and medications.
The method demonstrated robustness, stability, and conciseness in derived phenotypes.
A reduced set of 40 Limestone-derived phenotypes outperformed 640 original features in a heart failure prediction task, achieving an AUC of 0.720.
82% of the top 50 automatically extracted candidates were confirmed as clinically meaningful by a medical expert.

Conclusions:

Limestone offers an effective, unsupervised approach for EHR phenotyping, reducing reliance on manual supervision.
The method facilitates the discovery of clinically relevant patient phenotypes from complex EHR data.
Limestone shows significant potential for advancing data-driven clinical research and patient management.