Detecting Phosphorylation Determined Active Protein Interaction Network during Cancer Development by Robust Network Component Analysis

摘要: Motivation: In recent disease study, many key pathogen genes/proteins are found to have not significant differential expressions, and thus, they tend be disregarded in conventional expression analysis or network analysis. Meanwhile, the activity dry-experiment rather than wet-experiment been proposed effectively estimate actual regulation power of such important biomolecules, e.g. transcriptional factors. But, it is still unknown what how a hidden factor (e.g. phosphorylation) determines this kind virtual as [1]. Especially, for cancer development emergent reconstruct active protein interaction detect underlying phosphorylation pattern dynamic manner [2-7]. Methods: Based on c-Myc mouse model liver cancer, we first collected data at several developmental time points. Then, constructed rough background by conditional mutual information. Next, improved component its robustness, used advanced approach RNCA (Robust Network Component Analysis) time-dependent networks target different times simultaneously. Finally, considering experiment-qualities data, canonical correlation maximal between group proteins module), which could reveal sub-network determinate phosphorylation. Results: preliminary evaluated robustness comparing with other methods. And real biological rewired during development, corresponding proteins, their drivers This work can further early diagnosis diseases edge biomarkers [1-2], [3-4] dynamical [5-7].