Research On An Integrated Analysis Platform For Metabolomics Data

BACKGROUND AND OBJECTIVEWith the rapid development of metabolomics,a large number of data analysis methods and tools have been developed and widely used.Currently,the integrated analysis methods of metabolomics and other omics and the multi-functional integrated platform are the research hotspots.In order to mine the biomedical information in the massive data more scientifically and effectively,it is necessary to constantly improve the existing methods,create new methods,and develop a platform with complete functions,stable performance and convenient use,which can provide bioinformatics support for the mechanism and application research based on metabolomics.METHODSFirst,the methods of metabolomics data analysis were sorted out and compared,and then an integrated analysis platform was constructed by using these optimal methods.Based on multiple real datasets,its performance was comprehensively evaluated by comparing it with other platforms.Meanwhile,a new method for the detection of "metabolome-microbiome" inter-correlation was developed by combining approaches of metabolome and microbiome data transformation,confounding factors elimination,maximum information coefficient,and multiple regression analysis.A series of simulation datasets and real datasets were used to verify its validity and accuracy.Compared with other correlation methods,the merits of this method were evaluated comprehensively.Finally,the functions and performance of the platform were further improved by incorporating new methods,expanding the knowledge base,expanding the operating platforms,and improving the interactive experience.RESULTSThe integration platform for metabolomics data mining(IP4M,V2.0)is developedwith comprehensive functions,stable performance and ease to use.IP4 M has a total of 57 functions,which are divided into 8 modules,covering various links of metabonomics data mining,including peak identification,data preprocessing,difference/correlation/clustering analysis,machine learning for diagnostic models,pathway and enrichment analysis,etc.Compared with other comprehensive software,its unique sthengths are as follows:(1)A new method for metabolome-microbiome association analysis(GRa MM)and a new tool(3Mcor).(2)A series of ratio variables can be generated based on the self-built metabolic reaction database.These extended variables can indirectly reflect the activity of metabolic enzymes and reactions.(3)The knowledge base of the pathway analysis module contains about 15,000 compounds,1,600 pathways and 7,000 entries related to metabolic diseases.It also provides a variety of analytical algorithums.(4)It provides two modes of refinement and workflow,including Windows,Linux and Mac OS versions,which can meet the different users' needs.CONCLUSIONSIP4M has comprehensive functions,rich results and low threshold to use,and provides strong support for one-step analysis of metabolomics data.In the future,the platform will support more types of raw data and incorporate more methods about mulit-omics integration and network analysis.