Introduce an online platform for multi omics integration and network visual analysis

Gu He health

In order to make the research more systematic and comprehensive , More and more researchers are pursuing Multiomics background Explain the molecular data .

OmicsNet (www.omicsnet.ca) Born according to , This is an online platform , Allow users Easily build 、 Visualize and analyze multiomics Networks .

OmicsNet 2.0 It has been upgraded on the original basis , Three aspects have been improved ：

(1) Visual analysis 2D Graphic layout Options Added 11 individual , And a novel 3D modular .

(2) Three new omics types are supported ：

• From the study of genetic variation SNP list
• Species abundance matrix from microbiome analysis
• Liquid chromatography from non targeted metabonomics - Mass spectrum peak (LC-MS)

(3) The analysis will At the same time output The use of R package OmicsNetR And historical orders , And generate a shared link in Web page viewing and interaction operation , So as to improve the repeatability of the study .

Used in the study OmicsNet 2.0 For inflammatory bowel disease (IBD) The data were analyzed by multinomics .

OmicsNet 2.0 Workflow

Pictured , Mainly for 4 A step ：

• Upload data （ Support 8 Types of data ）
• Select database （ Select according to the data type ）
• To build the network
• visualization

OmicsNet 2.0 Improvement

1、 Updated the molecular interaction database

Include PPI database (STRING,InnateDB and IntAct),TF-target database (TRRUST and JASPAR), miRNA-target database (TarBase and miRTarBase), Metabolic database (KEGG, Recon3 and AGORA).

2、 Three new omics types are supported

Support from Genetic variation Studied SNP matrix 、 come from Non targeted metabonomics Of LC-MS Feng he comes from Species abundance matrix data for microbiome analysis Input and interaction analysis of .

For those who are right Mutations that affect gene regulation Interested users , Can be based on ADmiRE and SNP2TFBS take SNPs Mapping to miRNAs or TF Binding sites . The resulting network can be accessed through proteins、miRNAs or TFs Further expansion , With Understand the potential impact .

For metabonomic analysis , Use recently released NetID Algorithm to LC-MS peak (m/z, retention time, intensity and p-value) Data processing , can Choose three databases (KEGG, PubChemLite_BioPathway and HMDB) Annotate .

Use Rcpp/C++ The engine rewrites the core algorithm , Make it Increased speed 10 More than times . Use lpsymphony Package further optimizes integer linear programming .

Directly upload the species abundance matrix , Bayesian logic model can be used Predict potential metabolites , This model uses more than  6000  A high quality genome scale metabolic model Training .

3、 Open source code

to open up R package OmicsNetR（https://github.com/xia-lab/OmicsNetR）, Displays the R command . The results can be output as Web links , The period of validity A month .

4、 Reduce false positives

For data from humans and mice , Provides a ENCODE、 genotype - Tissue expression (GTEx) Or human protein map (HPA) Of Filters for gene expression data , These filters help researchers Focus on biological interactions , And reduce false positives .

5、 Support 2D and 3D Network Visualization

2D Module can Support 11 A graphic layout ,3D Module can Support interaction . Graphing is mainly based on igraph Bao He graphlayouts package .

Use OmicsNet 2.0 On going IBD Multigroup research

chart A by 2D View , chart B by 3D View

The species abundance matrix will be obtained ,SNP Matrix and LC-MS Peak data Upload to OmicsNet 2.0 platform .

utilize AGORA database Predict potential microbial metabolites (potential score:0.9).

utilize PhenoScanner Yes, based on eQTLs Of genes SNP location .

utilize KEGG database Yes LC-MS Peak to comment .

By adding Metabolites - Protein interactions , from SNPs and LC-MS The individual network generated by the peak is further expanded , Such three networks can be in Metabonomic layer merging .

p-value Filter ：(cut-off: 0.2) To exclude p Worth more LC-MS Node contributed by peak value . Its network connection is based on its related metabolites 、 Genes or proteins .

Pictured A, Generated subnet 1 contain Six types of nodes （ species 、SNPs、 Metabolites 、SNPs Or metabolite related genes / protein ）, The results show that glutathione It's the host - In microbial interactions Important interaction points .

It is speculated that several microorganisms including E. coli will produce glutathione, Two SNPs(rs3197999 and rs1428554) By encoding glutathione peroxidase (GPx) Genes GPX1 and GPX3 Related to metabolites .

Corresponding to this is , It has been proved in previous studies that E. coli and related species stay IBD The overexpression in may be due to their Can better produce glutathione To resist oxidative stress .

chart B Of 3D Layered networks provide a Intuitive multi group integration perspective , Highlights Connecting the microbiome to host genetics Of glutathione Flow path of .

OmicsNet 2.0 Comparison with other tools

The following table , Compare OmicsNet 2.0 With others based on web Of the multigroup integration tool Key features .

‘√’ Express support ,‘-’ Indicates missing ,“+” Indicates quantitative evaluation (“+” The more, the better )

The official website and main features of each tool are attached ：

PaintOmics

(http://www.paintomics.org)

Focus on Vision Present on Exploratory analysis of multiomics data , Including transcriptomics , Metabonomics , Epigenomics ,miRNA And transcription factors , Map it to KEGG Channel joint analysis .

MergeOmics

(http://mergeomics.research.idre.ucla.edu)

Integrated from Summary statistics of association studies at the individual omics level , And different functions Genomics data , To gain insights into the mechanism , Recently, we have also added multi - omics information Drug repositioning (drug repositioning).

OmicsAnalyst

(https://www.omicsanalyst.ca)

Based on the input data , utilize Multivariate statistics 、 Correlation analysis and clustering methods , Combined with the Internet 、 Heat map and scatter diagram are used for multinomial analysis .

Arena3D

(https://www.arena3d.org)

Good at using... Based 3D The hierarchical layout of the multi-layer network Interactive visualization , It is applicable to multi group network data .

NedRex

(https://nedrex.net)

It's a Cytoscape plug-in unit , Focus on Disease module identification and Drug reuse , Using various module identification and prioritization algorithms .

MetScape

（http://metscape.ncibi.org）

It's a Cytoscape plug-in unit , Through the construction and analysis of different types of enzymes 、 Metabolites and / Or a network of reactions , Focus on the integration and visualization of gene expression and metabonomics data .

OmicsNet

(https://www.omicsnet.ca)

By integrating multiple molecular interaction databases with powerful 2D/3D Visual network analysis , about Multiomics integrated analysis and Interpretation of the results Yes Better theoretical basis , Lack of the above tools Added .

junction language

OmicsNet 2.0 It's a Web-based Multi group study Analysis platform , Support 2D and 3D Network visualization exploration .

stay 1.0 The version emphasizes that it is based on web Of 3D The Internet visualization .

stay 2.0 In the version , The visual analysis system is further improved , Added a Full function 2D Network visualization system , And support Three new omics data inputs (SNPs、 Microbial taxa and LC-MS peak ), Current bioinformatics tools do not support these data well .

In addition to enrichment analysis 、 Search module and shortest path analysis , Users can also use Restart the random walk algorithm (Random Walk with Restart) Search for candidate disease markers . Last ,2.0 Version by releasing the underlying R Code and shared links , improvement With tools Reproducibility and transparency .

Use IBD Case studies of multiomics data show that ,OmicsNet 2.0 Can be revealed with original and subsequent publications as well as IBD A meaningful pattern consistent with the literature 、 Relationships and functions . All in all ,OmicsNet 2.0 It solves some requirements for the analysis of multiomics data .

reference ：

Zhou G, Pang Z, Lu Y, Ewald J, Xia J. OmicsNet 2.0: a web-based platform for multi-omics integration and network visual analytics. Nucleic Acids Res. 2022 May 26:gkac376. doi: 10.1093/nar/gkac376. Epub ahead of print. PMID: 35639733.