As part of the ProteomicsDB project, this project will lead to an extended data model allowing the integration of additional information such as signaling pathways and drug meta data into ProteomicsDB. The integration of phenotypic drug sensitivity data from different sources will enable the implementation of various machine learning approaches in ProteomicsDB, in order to discover biomarkers predicting drug sensitivity and resistance.
Furthermore, the integration of additional information such as protein structures, signaling pathways, protein-protein interaction networks and drug sensitivity data will enable us to learn and, ultimately, predict drug sensitivity and resistance given available molecular signatures such as mutations, transcriptome, proteome and PTM expression data of (unknown/unclassified) cell lines or patients.
The high impact output of this project is a platform and a set of algorithms for evidence-based medicine using different molecular signatures of a biological sample to predict and guide treatment decisions.
The extension of ProteomicsDB to any organism, here with a strong focus on mice, rats and mini pigs, will enable the comparison of expression values of proteins in toxicity studies between preclinical models and humans in order to be able to better translate pharmacokinetic and pharmacodynamic observations to the human situation.