Exploring the structure and sequence diversity in human kinome

Abstract

The report outlines the work which was done in the summer term, including two primary tasks, namely, creating a web site with data on human kinases and integrating data on various kinases databases. Web development work was done where a new web version of the web kinase.com was developed. The new frontend was developed in the latest and user-friendly HTML, CSS, JavaScript, and React.js themes. The site has enhanced search tools, data view visualizations, and 3D protein structures using the NGL Viewer. It uses Node.js and Express.js to create the backend together with MongoDB that enables the system to process data requests and deliver protein structures executing the API endpoints. Besides web development, integration of a comprehensive kinase database was done. Several kinase-specific databases (KinBase, EKPD, KLIFS, KLSD, Kincore, and ChEMBL) were examined to learn what information they contain: total number of kinases, the collection of species, classification schemes, and kind of data that are being provided. A merging pipe-line was established with UniProt IDs to aggregate the data into a single completed one. It led to the compilation of 719 proteins of 492 unique human kinases, amalgamated out of 11 data-sets, including UniProtDB, KINHUB, KLIFS, and KinBase. In order to introduce the data and interpret it, bar charts, UpSet plots, and heatmaps were made, which demonstrates the size of each dataset and their overlaps. KEGG pathway analysis, enrichment, and kinase expression profiling in human tissues based on GTex was also conducted further to determine significant biological pathways and particular tissues in which active kinases are expressed.