Josh F. Peterson

Diagnostic Accuracy of Large Language Models for Rare Diseases: A Systematic Review and Meta-Analysis

This paper evaluates the diagnostic accuracy of large language models for rare diseases through a systematic review and meta-analysis, finding substantial variation in performance across benchmarks and high risk of bias in all studies.

Implementing Integrated Genomic Risk Assessments for Breast Cancer: Lessons Learned from the eMERGE Study

The paper develops a pipeline for integrated breast cancer risk assessment using polygenic risk scores and other factors in the eMERGE study, showing high concordance between API and UI results.

Risk factors affecting polygenic score performance across diverse cohorts

The paper examines how covariates affect polygenic score performance across European and African cohorts, finding significant differences in R2 and interaction effects, and proposes methods to improve model performance.

Expansion of clinical genetic testing since the completion of the human genome project

This study quantifies the expansion of clinical genetic testing at Vanderbilt University Medical Center from 2002 to 2022, showing increased testing rates, broader disease coverage, and a growing number of variants of uncertain significance.

The Phenotype-Genotype Reference Map: Improving biobank data science through replication.

The paper introduces a phenotype-genotype reference map for replicating GWAS results in biobank data and demonstrates its utility in detecting data corruption and assessing reproducibility.

Selection, optimization, and validation of ten chronic disease polygenic risk scores for clinical implementation in diverse populations

The paper selects and validates ten polygenic risk scores for clinical use in diverse populations, focusing on chronic diseases and addressing disparities in predictive performance.

Phenotypic convergence: a novel phenomenon in the diagnostic process of Mendelian genetic disorders

The paper develops an EHR-based model to study the diagnostic trajectory of Mendelian disorders and finds that phenotypic features are often documented after clinical suspicion, a phenomenon termed phenotypic convergence.

A scalable EHR-based approach for phenotype discovery and variant interpretation for hereditary cancer genes

This study uses EHR data to identify gene-phenotype associations and develop phenotype risk scores for variant interpretation in hereditary cancer genes.