Abstract
BackgroundProtein kinases are a large and diverse family of enzymes that are genomically altered in many human cancers. Targeted cancer genome sequencing efforts have unveiled the mutational profiles of protein kinase genes from many different cancer types. While mutational data on protein kinases is currently catalogued in various databases, integration of mutation data with other forms of data on protein kinases such as sequence, structure, function and pathway is necessary to identify and characterize key cancer causing mutations. Integrative analysis of protein kinase data, however, is a challenge because of the disparate nature of protein kinase data sources and data formats.ResultsHere, we describe ProKinO, a protein kinase-specific ontology, which provides a controlled vocabulary of terms, their hierarchy, and relationships unifying sequence, structure, function, mutation and pathway information on protein kinases. The conceptual representation of such diverse forms of information in one place not only allows rapid discovery of significant information related to a specific protein kinase, but also enables large-scale integrative analysis of protein kinase data in ways not possible through other kinase-specific resources. We have performed several integrative analyses of ProKinO data and, as an example, found that a large number of somatic mutations (∼288 distinct mutations) associated with the haematopoietic neoplasm cancer type map to only 8 kinases in the human kinome. This is in contrast to glioma, where the mutations are spread over 82 distinct kinases. We also provide examples of how ontology-based data analysis can be used to generate testable hypotheses regarding cancer mutations.ConclusionWe present an integrated framework for large-scale integrative analysis of protein kinase data. Navigation and analysis of ontology data can be performed using the ontology browser available at: http://vulcan.cs.uga.edu/prokino.
Highlights
Cancer is caused by an accumulation of mutations, often in a subset of genes that confer survival and growth advantage
Targeted sequencing of 518 protein kinase exons encoded in the human genome has revealed hundreds of mutations in the protein kinase domain [2]. These mutations are currently catalogued in various databases [3,4,5], identification and experimental characterization of key cancer-causing mutations is essential for developing new therapies for cancer
An aggregate query ‘‘counts of kinase mutations in various cancer types’’, for example, revealed that the mutations associated with haematopoietic neoplasm (288 distinct mutations) primarily target only 8 kinases in the human kinome, compared to glioma, where the mutations are spread over 82 distinct kinases
Summary
We describe ProKinO, a protein kinase-specific ontology, which provides a controlled vocabulary of terms, their hierarchy, and relationships unifying sequence, structure, function, mutation and pathway information on protein kinases. We have performed several integrative analyses of ProKinO data and, as an example, found that a large number of somatic mutations (,288 distinct mutations) associated with the haematopoietic neoplasm cancer type map to only 8 kinases in the human kinome. This is in contrast to glioma, where the mutations are spread over 82 distinct kinases. We provide examples of how ontology-based data analysis can be used to generate testable hypotheses regarding cancer mutations
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
