Phosphorylated protein chip combined with artificial intelligence tools for precise drug screening

Abstract

We have developed a protein array system, named “Phospho-Totum”, which reproduces the phosphorylation state of a sample on the array. The protein array contains 1 471 proteins from 173 known signaling pathways. According to the activation degrees of tyrosine kinases in the sample, the corresponding groups of substrate proteins on the array are phosphorylated under the same conditions. In addition to the measured phosphorylation levels of the 1 471 substrates, we developed and performed the AI-assisted tools to further characterize the phosphorylation state and estimate pathway activation, tyrosine kinase activation, and a list of kinase inhibitors that produce phosphorylation states similar to that of the sample. The Phospho-Totum system, which seamlessly links and interrogates the measurements and analyses, has the potential to not only elucidate pathophysiological mechanisms in diseases, by reproducing the phosphorylation state of samples, but also be useful for drug discovery, particularly for screening targeted kinases for potential drug kinase inhibitors.