The Construction Of Human Phosphorylation Mediated Signaling Transduction Network And Analysis Platform

Protein function network constitutes an indispensable part of high throughput data analysis.The signaling network has important functional significance,which has attracted considerable attention in biomedical research areas.The key factor in the regulation of signaling network is the dynamic protein phosphorylation,which is due to the switch of kinases.Most studies of phosphorylation only took the kinase-substrate regulatory relationship into account.However,the construction of phosphorylation mediated whole signaling transduction network is in lack,which should contain entire signaling transduction cascades(i.e.,ligand-receptor,GPCR-G protein-effector,kinase-substrate,phosphatase-substrate,and transcription factor-target gene).Such a network can provide the background for molecular functional analysis.Further,various levels of omics data can be mapped into the network,and the comparative study of functional networks in different states can be elaborated,such as cancer state versus normal state.In this work we aim to construct human phosphorylation mediated signaling transduction network,and build analysis platform for proteomics and cancer researchers.The main steps include: signaling transferring pairs of each cascade were collected,particularly phosphorylation related protein nodes,to construct the database named PhoSigNet;cancer-related information of protein nodes in the network were collected as multi-level annotation of PhoSigNet,including proteins with differential expression and variations in human cancer;then a time series data analysis tool was developed in PhoSigNet,because lots of studies carry out experiments on several time points to measure the dynamic regulation of phosphorylation;finally a cancer application platform was developed according to cancer-related multi-level annotation,to discover cancer-triggering receptors with enrichment analysis.Meanwhile,given the importance of proteins with differential expression and variations in human cancer,we updated the database of differentially expressed proteins in human cancers(dbDEPC),and human cancer proteome variation database(CanProVar)through integration of all proteins with the cancer-related information.Through integrating three types of information,including phosphorylation sites,kinase-substrate relationships,and upstream and downstream signaling interactions,the constructed signaling network contains 11,976 experimentally validated directed edges and 216,871 phosphorylation sites.Moreover,the annotation of differentially expressed proteins contains over 4000 protein entries,curated from 331 experiments across 20 types of human cancers;the annotation of human cancer variation sites contains 70,262 crVARs in 23,816 proteins.Then,we developed a clustering analysis tool,named ExpCluster,to facilitate time series analysis of proteomics data;a cancer application platform,named CanReceptor,to fulfill cancer-triggering receptor analysis.This project provided a comprehensive platform for studies of phosphorylation sites,protein kinases and signaling paths.It would be very useful in the mechanism and function studies of numerous phosphoproteomics experiments.At the same time,protein annotation with differential expression and variations in human cancer would benefit cancer mechanism research,and laid a foundation for future target screening and drug design in cancer-related signaling pathways.