Damian P. Smedley

This paper describes an AI semantics workflow using LLM agents to integrate multimodal biomedical data through ontologies and knowledge graphs.

The paper identifies pathogenic variants in 15.8% of CP patients and finds only 16 of 60 genes have significant CP association.

The paper uses blood-based RNA-Seq on 5,412 individuals with rare diseases to identify new candidate diagnoses and shows its utility in prioritizing genetic variants.

This study benchmarks large language models against a traditional diagnostic tool for genetic disease diagnosis using 5,213 case reports, finding LLMs lag behind in accuracy.

The paper introduces GPSEA, a tool using GA4GH Phenopackets to analyze genotype-phenotype correlations in Mendelian disorders, identifying 225 significant associations across 85 cohorts.

The paper identifies MAEA as a critical regulator of DNA repair and replication, with loss of function causing developmental delay and neurodevelopmental deficits.

The paper evaluates two self-phenotyping tools, showing that GenomeConnect provides more accurate phenotype profiles while Phenotypr offers greater precision and user preference for its autocomplete format.

The paper describes the International Mouse Phenotyping Consortium's data portal, its expansion to over 100 million data points, and improvements to its web interface and data access tools.

PhEval provides a standardized framework to benchmark phenotype-driven variant and gene prioritisation algorithms using real-world patient data and controlled test corpora.

The paper introduces uPheno, a framework for cross-species phenomics integration by harmonizing phenotype data across species and ontologies.

This study analyzes genetic causes of syndromic and non-syndromic hearing loss in 5,488 individuals from the 100,000 Genomes Project, identifying 273 genes and a 27.5% diagnostic yield.

The paper analyzes gene intolerance to loss-of-function variation in a rare disease cohort to identify patterns in gene expression, sequence features, and disease associations across FUSIL categories.

The paper presents a large-scale corpus of standardized phenopackets for genomic diagnostics and discovery, curated from literature and useful for software testing and genotype-phenotype analysis.

The paper evaluates variant prioritization methods for rare disease diagnosis using data from the Rare Genomes Project, finding that top models identify causal variants in most solved families and uncover potential candidates in unsolved ones.

This paper curates human essential genes based on lethality in Mendelian disorders and explores their relationships with constraint metrics and viability data.

The paper identifies 88 novel rare disease-gene associations using a burden analysis framework on whole genome sequencing data from the 100,000 Genomes Project, with potential clinical impact for undiagnosed cases.

This study identifies TMEM63B variants in patients with developmental and epileptic encephalopathy and neurodegeneration, showing altered cation currents and hematological abnormalities.

This paper introduces OBA, a formalized ontology for biological attributes to facilitate data integration across phenotypic, genetic, and environmental data.

The paper introduces the GA4GH Phenopacket schema to standardize the exchange of phenotypic data for precision medicine applications.

This study uses whole-genome sequencing to identify regulatory and novel genic variants contributing to childhood cardiomyopathy.

SQUIRLS prioritizes splice variants by calculating interpretable features and achieving higher accuracy and speed than existing methods in simulated exomes.

SvAnna predicts pathogenicity of structural variants in long-read sequencing data by integrating transcript and regulatory annotations with clinical phenotype data, achieving high accuracy in prioritizing pathogenic variants.

The paper identifies gene-specific mutational signatures caused by replication errors and endogenous DNA damage using CRISPR-Cas9 knockouts and whole-genome sequencing.

The paper evaluates how well machine-learning models capture ontology relationships and suggests ways to improve curation using these models.

The paper reduces the dimensionality of phenotypic data from 53,000 mouse models to visualize gene function diversity and highlights challenges in embedding sparse phenotypic data.

OpenStats improves speed and scalability for high-throughput phenotypic data analysis, showing significant performance gains in the International Mouse Phenotyping Consortium.

The paper introduces LIRICAL, a likelihood ratio-based framework for genomic diagnostics that improves accuracy and interpretability in phenotype-driven rare disease diagnosis.

The paper integrates gene essentiality data from human and mouse screens to classify genes into FUSIL categories, identifying developmental lethal genes associated with rare diseases.

The paper introduces a method called soft windowing to improve analysis of high-throughput phenotyping data by reducing false positives and increasing significant results.

The paper uses whole genome sequencing to identify disease-causing variants in non-coding regions of genes, confirming diagnoses of Cystic Fibrosis and Primary Ciliary Dyskinesia.