ProKinO

ProKinO integrates diverse protein kinase data into a knowledge graph and ontology to enable analysis of sequence, structure, mutation, expression, and ligand interaction relationships for discovery and hypothesis generation.

Key Features:

• Expanded Knowledge Base: The Protein Kinase Ontology has been expanded with classes and relationships capturing kinase ligand binding sites, expression patterns, functional features, and drug interactions.
• Graph mining and SPARQL queries: Aggregate SPARQL queries and graph-mining techniques are applied to the ontology to extract patterns and associations from integrated kinase data.
• Mutation and cancer findings: Data mining of the ontology has identified frequent mutations, including p21-activated protein kinase 5 (PAK5) mutations observed in human cancers.
• ProtVista integration: ProtVista is integrated to map kinase sequence annotations onto 3D structures and Alphafold models for combined sequence–structure annotation analysis.
• Drug binding residue identification: The ontology-backed analyses identify common ligand/drug binding residues across PAK family kinases relevant to drug discovery.

Scientific Applications:

• Understudied kinase characterization: Enables generation of novel insights into understudied members of the protein kinase family by linking sequence, structure, expression, and mutation data.
• Oncogenic mutation discovery: Supports identification of oncogenic and recurrent mutations such as those in PAK5 across human cancers.
• Expression pattern analysis in cancer: Facilitates analysis of abnormal expression patterns in cancers, including acute myeloid leukemia.
• Drug target and binding-site analysis: Informs identification of potential therapeutic targets and common ligand/drug binding residues in PAK family kinases for drug discovery.

Methodology:

Integration of diverse datasets into the Protein Kinase Ontology; systematic expansion of ontology classes and relationships; application of graph-mining techniques and aggregate SPARQL queries; integration of ProtVista for mapping annotations to 3D structures and Alphafold models.