Application Of Multi-omics Data Analysis Method Based On Biological Network To Study The Mechanism Of Action Of Compounds

With the development of high-throughput experimental technology,multi-omics data has exploded,providing data foundation for a comprehensive understanding of complex diseases and physiological processes.The biological network-based method provided a knowledge framework for the integration and interpretation of multi-omics data,and becomes an important means for multi-omics data mining.By mapping multi-omics data to different physiological systems in the form of networks,it was expected to generate new breakthroughs in key questions of complex diseases and biological processes based on the network level.This study summarizes the application of multi-omics approaches and network analysis in drug discovery and the impact of compounds on complex physiological processes in human body.Finally,the effects of the compounds were further investigated.In the first part of the work,we constructed a computational framework based on placental gene networks through analysis methods for multiomics data and biological networks.The function of multiple placental transporters in mediating compound transport across the placenta was systematically considered through this framework.Each compound was generated a probability score for crossing the placental barrier,which comprehensively assessed the compound’s ability to cross the placental barrier.Finally,the transport processes of several typical environmental compounds in the placental barrier were analyzed by molecular dynamics simulation,using dipalmitoyl phosphatidylcholine(DPPC)bilayers as a simple placental barrier model.The probability score of crossing the placental barrier was generated for 307 compounds based on the calculation framework of the placental gene network,and the ability of the compounds to cross the placental barrier was divided into high,medium and low potential intervals according to the distribution of probability scores.Subsequently,the probability scores of 259 environmental compounds crossing the placental barrier were also calculated,and the interaction of compounds with different probability scores with the DPPC bilayer membrane model was analyzed by molecular dynamics simulation.The results showed that there were also obvious differences in the effects of compounds distributed in different probability intervals on DPPC bilayer membranes.Compounds with higher probability scores were more likely to enter the DPPC bilayer membrane.Compounds with lower probability scores did not easily enter the DPPC bilayer membrane.The above results suggested that a computational framework based on the placental gene network can be used for risk assessment of unknown compounds crossing the placenta.In the second part of the work,the estrogenic interference effects of G protein-coupled estrogen receptor 1(GPER)mediated Triphenyl phosphate(TPP)were analyzed by combining transcriptomics and proteomics methods and refining its adverse outcome pathway(AOP)framework through molecular dynamics simulation and fluorescence analysis.Transcriptomics and proteomic studies of SKBr3 cells exposed to TPP revealed that 2048 differentially expressed genes(DEGs)and 591 differentially expressed proteins(DEPs)were identified as co-existing 64 DEGs/DEPs,of which 13 genes and proteins were up-regulated and 39 were down-regulated.The enrichment results of these 52 co-expression DEGs/DEPs pathways showed that GPER-related pathways such as PI3K-Akt signaling pathway,MAPK signaling pathway,Erb B signaling pathway,Wnt signaling pathway,EGFR tyrosine kinase inhibitor resistance and estrogen signaling pathway were enriched.Molecular dynamics simulation results and fluorescence analysis showed that TPP binding to GPER could change from an inactive state to an active state,and binding to GPER was determined as the molecular initiation event leading to the estrogenic interference effect of TPP.The results of omics analysis further complemented the relationship between molecular initiation events and key events of AOP.Therefore,the AOP framework based on multi-omics results showed that TPP activation of GPER may affect cell proliferation,metastasis and apoptosis,regulate gene transcription and kinase activity,lead to abnormal estrogen-dependent cell processes such as immune function and cancer,and ultimately lead to the occurrence of estrogen interference effects.Multi-omics studies provided an effective method for further refining the AOP framework for TPP-induced estrogen interference.In the third part of the work,we integrated the protein-protein similarity network,compound-compound similarity network,and compound-target interactions(CTIs)network into one network,and re-predicted CTIs by constructing a compound-target interactions network(CTI Net)associated with Alzheimer’s disease(AD)by Laplace regularized least squares method.The screening results based on CTI Net showed that daphnetin was identified as a high potential compound for the treatment of AD.Further,transcriptomic analysis was used to investigate the mechanism of action of daphnetin in the treatment of AD.The results showed that daphnetin mainly affected the mitochondrial pathway and the neurofibrillary tangles pathway composed of highly phosphorylated microtubule-associated protein(Tau)in brain cells,so as to achieve the therapeutic effect of AD.The methods based on network and omics analysis were helpful for screening high potential compounds for the treatment of AD,and provided a reference for revealing the mechanism of action of candidate compounds for the treatment of AD at the molecular level.