David Ochoa

This paper introduces Karenina, a framework that evaluates biomedical AI using expert knowledge across multiple dimensions beyond traditional scoring.

VIDRA models genetic dose-response relationships to predict drug mechanisms, therapeutic potential, and adverse events using phenotypically associated variants.

This paper analyzes 100,526 GWAS to identify pleiotropic genes and their therapeutic implications, finding that highly pleiotropic genes are associated with safety liabilities and that combining PAV support with intermediate pleiotropy improves therapeutic success.

This study systematically analyzes chemical probe literature to identify target-disease associations, revealing their utility in early therapeutic discovery and highlighting opportunities for repurposing.

The paper maps cell-type-specific eQTLs in IBD patients to identify effector genes and mechanisms linking genetic risk to disease.

The paper introduces a framework for extracting biological evidence from literature to aid drug discovery, identifying millions of associations between targets, diseases, and drugs.

The paper introduces 'Associations on the Fly' (AOTF), a new feature in the Open Targets Platform that allows users to dynamically adjust evidence weights for target-disease associations.

The paper identifies 995 genes with shared genetic links to age-related diseases using GWAS data and presents them in a web tool for drug target prioritization.

The paper uses network-based expansion of GWAS results to identify pleiotropic gene modules linked to multiple traits and demonstrates their application in studying inflammatory bowel disease.

The paper analyzes reasons for clinical trial failures using NLP and shows that genetic evidence improves trial success chances.

The paper presents an open resource for fine-mapping and gene prioritization at GWAS loci, identifying causal variants and genes and improving drug target identification.

The paper systematically maps the functional role of phosphorylation by analyzing phospho-deficient mutants in yeast and linking yeast phenotypes to human functional relevance.

The paper creates a comprehensive human phosphoproteome by integrating 6,801 phospho-enriched mass spectrometry experiments and prioritizes functional phosphosites using machine learning.

The paper characterizes interaction-mediated control of protein abundance using multi-omics data from cancer samples and finds that 42% of proteins show post-transcriptional regulation.

The paper identifies copy number aberrations that alter kinase signaling and highlights potential therapeutic targets by integrating genomic and proteomic data across cancer samples.

The paper presents a dataset of 28,000 annotations describing how amino acid mutations affect protein interactions, available through the IntAct website.

The paper benchmarks kinase activity inference methods using phosphoproteomic data, finding the GSEA method performs best with an average AUC of 0.722.

This study integrates metabolomics, phosphoproteomics, and transcriptomics data in yeast to identify regulatory mechanisms linking signaling kinases to metabolic changes.

This paper constructs a kinase regulation atlas using phosphopeptide quantifications to identify co-regulated kinases and their functional relationships, revealing key mediators in stem cell differentiation and Salmonella infection.