Using GPT-4 to annotate the severity of all phenotypic abnormalities within the human phenotype ontology - Report - MDSpire

Using GPT-4 to annotate the severity of all phenotypic abnormalities within the human phenotype ontology

  • By

  • Kitty B. Murphy

  • Brian M. Schilder

  • Nathan G. Skene

  • May 21, 2026

  • 0 min

Share

Clinical Report: Employing GPT-4 for the Assessment of Severity in Phenotypic Abnormalities

Overview

{'true_positive_recall': 'Clarify the range of true positive recall rates and provide context on what constitutes high accuracy.'}

Background

{'AI_enhancement': "Include specific examples of how AI technologies like GPT-4 can enhance the HPO's utility."}

Data Highlights

{'mean_recall': 'Ensure consistency in reporting the mean recall rate and its significance.'}

Key Findings

{'severity_scoring': 'Expand on how the severity scoring system integrates clinical characteristics and its implications.'}

Clinical Implications

{'limitations': 'Discuss potential limitations or challenges in integrating AI into clinical workflows.'}

Conclusion

{'impact': 'Reiterate the potential impact of this approach on clinical practice and research, emphasizing future directions.'}

Related Resources & Content

  1. Acta Neuropathologica, 2020 -- Comprehensive Analysis of Pathway-Specific Polygenic Risk and Transcriptomic Networks Reveals New Functional Pathways in Parkinson's Disease
  2. Acta Neuropathologica, 2021 -- Modeling Latent Traits of Tau Pathology in Progressive Supranuclear Palsy
  3. Brain, 2025 -- Neuroanatomical Subtypes of Primary Progressive Aphasia Identified Through Data Analysis
  4. Brain, 2025 -- Biallelic Variants in PTPMT1 Impair Cardiolipin Metabolism, Resulting in a Neurodevelopmental Disorder
  5. GA4GH -- Phenopackets
  6. An artificial intelligence framework for end-to-end rare disease phenotyping from clinical notes using large language models
  7. Large language models in real-world clinical workflows: a systematic review of applications and implementation - PMC
  8. Phenopackets – GA4GH
  9. An artificial intelligence framework for end-to-end rare disease phenotyping from clinical notes using large language models
  10. Large language models in real-world clinical workflows: a systematic review of applications and implementation - PMC

Original Source(s)

Using GPT-4 to annotate the severity of all phenotypic abnormalities within the human phenotype ontology

  • by Kitty B. Murphy, Brian M. Schilder, Nathan G. Skene

  • May 21, 2026

Related Content

Refining Single-Cell Metabolomics with Ion Mobility

Ion mobility-resolved mass cytometry expands metabolite coverage across thousands of single cells 

  • February 24, 2026

  • 2 min
Feasibility and implementation of a daily safety brief at a children's hospital-in-a-hospital

Surgical Approach to Glandular Urethral Reconstruction in a Case of Effmann Type I A1 Urethral Duplication with Glans Duplication