Imeta | Nannong shenqirong team released microbial network analysis and visualization R package ggclusternet

Liuyongxin Adam2022-06-23 18:10:56

Click on the blue words Pay attention to our

ggClusterNet: Contains a variety of microbial network mining algorithms based on module visual layout R package

936a9981035097f5777291e3413bfa54.png

https://doi.org/10.1002/imt2.32

SHORT COMMUNICATION

Volume1, Issue3

7a9efe2785a6e5add9de4838e439ffd2.png

●2022 year 6 month 13 Japan , Nannong shenqirong's team is in iMeta Published online entitled “ggClusterNet: An R package for microbiome network analysis and modularity-based multiple network layouts” The article .

●  This article developed a ggClusterNet Of R package , Exhibition Microbial network modular information , For microbial network data mining and cross domain network data mining .

●  First author : Wen Tao , Xiepenghao

●  Corresponding author : Liu Yongxin  , Yuan Jun (junyuan@njau.edu.cn)

●  joint authorship : Yangshengdie , Niuguoqing , Liuxiaoyu , Dingzhexu , Xue Chao , Shen Qirong

pick     want

Network analysis is gradually valued by ecologists and continues to be applied in the field of ecology , It is necessary to develop more powerful and convenient network analysis tools . therefore , We developed a program called ggClusterNet Of R package , For easier network data analysis, mining and visualization . stay ggClusterNet Dozens of network layout algorithms are designed in the package to better display the modular information of microbial network (randomClusterG, PolygonClus-terG, PolygonRrClusterG, ArtifCluster, randSNEClusterG, PolygonModsquar-eG, PolyRdmNotdCirG, model_Gephi.2, model_igraph, and model_maptree). For the convenience of researchers , Designed a variety of microbial network data mining functions , For example, correlation calculation (corMicor()), Network attribute calculation (net_properties()), Node attribute calculation (node_properties()), Random network calculation and comparison (random_Net_compate()). In order to mine the network more quickly , Further integrate these functions into two functions network.2() and corBionetwork(), It is used in microbial network data mining and cross domain network data mining respectively . at present ,ggClusterNet stay github(https://github.com/taowenmicro/ggClusterNet/),Gitee(https://gitee.com/wentaomicro/ggClusterNet) Open for use on . The complete description and examples are in wiki Pages can be read .

key word : Microbiome , Network analysis , R package , visualization

bright     spot

cbf9810cf8a9756e21cc7fcec91c107d.png

● ggClusterNet For mining microbial networks and cross domain networks , At the same time, it provides dozens of network layout algorithms based on modular display , Adapt to microbiome network mining

Video interpretation

Bilibili:https://www.bilibili.com/video/BV1ig411Q7cw/

Youtube:https://youtu.be/Kxi_pn4QJrU

Chinese translation 、PPT、 in / Download extended materials such as English video interpretation

Please visit the official website of the Journal :http://www.imeta.science/

Full text interpretation

Development and status of microbiome network analysis methods

Network analysis is not just exploring a set of projects ( node ) And the connection between them ( edge ) Mathematics of 、 Statistical and structural properties play an important role , It is also widely used to explore co-occurrence patterns among microbial taxa in complex communities . for example , Mabin et al. Found the interaction mode of microbiome in different environments based on network analysis , The importance of network analysis to microbiome community analysis is emphasized ;Yuan Through network analysis, et al. Found that climate warming enhanced the complexity and stability of microbial networks ;Franciska By using network analysis, et al. Found that drought would cause great disturbance to the soil bacterial network , But it has little effect on the fluctuation of fungal network .

The tools of microbiome network analysis mainly include web page analysis tools MENA(MENAP)、R software package (WGCNA、igraph、ggraph、SpiecEasi)、 Interactive biological software (Cytoscape and Gephi)、python Model (NetworkX and SparCC) etc. . The main tools for network related computing are :MENA、WGCNA、SpiecEasi etc. .MENA The correlation matrix of microbial network can be calculated , Based on random matrix theory (RMT) Method , It is robust to noise ;WGCNA It is used to construct a scale-free topology weighted network based on soft threshold , Weighting by other correlations outside the nodes is more biologically meaningful ;SpiecEasi The data transformation developed for the data is combined with the graphical model reasoning framework , Generated a set of calculation tools , It can effectively solve the robustness problem of small samples , Reducing false positives in correlation calculations . The main tools for network visualization are Cytoscape、Gephi and R package (igraph、ggraph etc. ).Cytoscape Not only can it provide a powerful visualization scheme , It also allows users to develop new functions through a variety of plug-ins , It can complete microbial network mining, etc ;Gephi Beautiful network graphics can be easily completed through a few steps .R Visualization tools in software packages (igraph、ggraph) You can also quickly show the network in the command line , And it has high repeatability . however , Although these tools have advantages in different aspects of network analysis , But it still can not meet our requirements for fast computing , Repeatability , In depth mining of network requirements . for example ,Cytoscape The steps are cumbersome , Not easy to repeat , and igraph and ggraph The layout of the is not beautiful .

at present , More researchers pay attention to the interaction between microbial network modules , To explore their functions . However , There are too many connections between various microbial species in the network , Moreover, there is a lack of appropriate visualization scheme and software to clearly show the relationship between modules , These tools are not enough to meet the needs of relevant researchers in practical use . In order to solve the increasing demand in network analysis , We go through R Language development R package ggClusterNet. It provides a fast way for network analysis 、 Repeatable and easy-to-use processes , And has a variety of powerful visual layout based on modularization .ggClusterNet It has the following characteristics :1) The network analysis process can be completed quickly ;2) The network analysis process can provide a variety of beautiful visual layouts ;3) The network analysis process can be completed with a small amount of code , Very repetitive .

ggClusterNet Introduction to development environment and process

ggClesterNet Is based on R Language development of network analysis process tools . among , The tools used in Network Computing mainly call cor() function (R package WGCNA),sparccboot() function (R package SpiecEasi) and corr.test() function (R package psych). Some network layouts refer to ggraph and sna package . The network attribute calculation tool mainly calls igraph Functions in the package (average.path.length()、edge_connectivity()、no.clusters()、centralization.closeness()、erdos.renyi.game() etc. ).ggClusterNet Software open source , And can be in Github Download the use (https://github.com/taowenmicro/ggClusterNet/), It can also be done through R Language command line tools for installation (devtools::install_github("taowenmicro/ggClusterNet")).

To enrich ggClusterNet Support for modular layouts , More than... Have been developed here 10 A network visualization layout function based on modularization , These functions are :randomClusterG、PolygonClusterG、PolygonRrClusterG、ArtifCluster、randSNEClusterG、PolygonModsquareG、PolyRdmNotdCirG、model_Gephi.2、model_igraph and model_maptree. These algorithms require correlation matrix and modular information of nodes as input , Can calculate various styles of network layout . These layout results can finally be used R Language ggplot2 Package for visualization .

debc22c1d59adccd57982c3f6e670081.png

chart 1. ggClusterNet Function and function introduction

ggClusterNet Use

●  ggClusterNet workflow

stay ggClusterNet in ,corMciro()( Correlation matrix calculation for microbial networks ) or corBiostripe()( It is used to calculate the correlation matrix of the binary network ) Is used to calculate the correlation matrix . More than ten network layout algorithms are designed to calculate the visual layout of correlation matrix , The calculation results are combined with ggplot2 Drawing . Separate use net_properties(),node_properties() and ZiPiPlot() To calculate network properties 、 Node properties , And calculate the function of nodes according to the module . Use random_Net_compate() Calculate the zero model and generate a random network , And compare with the network properties in microbial network . These functions are included in nerwork.2()( Microbiome network process ) or corBionetwork()( The flow of the binary network ), You can run the network analysis through a function . All in all ,ggClusterNet It can quickly complete the analysis of the whole microbiome and binary network , Including correlation calculation 、 Network Visualization 、 Network attribute calculation 、 Construction and comparison of node attributes and random networks .

1c1447a3d0a6533b7927f5ab57ff6651.png

chart 2. ggClusterNet Function usage and output results

●   Network layout

In order to better realize the visualization of microbial network and highlight the interaction between modules , Ten visual layout algorithms are developed . The following describes the functions of these algorithms .

1)randomClusterG: Put a single module ( Group ) The nodes of are all arranged in a ring , Then draw multiple modules into multiple circles with the same radius and randomly arrange these circles on the drawing panel .

2)PolygonClusterG: Put a single module ( Group ) The nodes of are arranged in a circle . These multiple modules are drawn as multiple circles of the same radius . Finally, these rings representing different modules are arranged on the polygon vertices with the same number of edges .

3)PolygonRrClusterG: Single module ( Group ) The nodes of are arranged in a ring . Draw the module into several circles with different radii according to the number of nodes in the module ( The more nodes , The larger the radius ). then , These circles are regularly arranged on the vertices of the polygon centered on the coordinate axis ( The number of edges equals the number of modules ).

4)ArtifCluster: Single module ( Group ) The nodes of are arranged in a circle with the same radius . By setting coordinate values , Arrange these circles manually .

5)randSNEClusterG: Single module ( Group ) Node of can call sna Visual layout calculation coordinates in the package , Different modules are randomly arranged on the drawing panel .

6)PolygonModsquareG: A single module ( Group ) The nodes of are arranged in a circle . Different modules are drawn as circles with different radii ( The more nodes , Larger radius ). These circles are arranged in one or more lines with a few parameters .
7)PolyRdmNotdCirG: According to the module information , The nodes are randomly distributed in several circles with different radii ( The more nodes , Larger radius ). These modules are regularly arranged on the vertices of the polygon centered on the origin of the coordinate axis ( The number of edges equals the number of modules ).

8)model_Gephi.2: All the nodes calculate the coordinates and arrange them into a circle . Reassign coordinates using microbial clustering results .

9)model_igraph: Use Fruchterman and Reingold The algorithm of , Project all nodes onto the coordinate axis according to the module classification .

10)model_maptree: Firstly, the network is analyzed by modularization , The nodes are grouped according to the degree of network modularity , Then it is used to calculate the coordinates . The relative position of nodes is calculated according to the algorithm developed by wangweixin et al , The algorithm attempts to map the same module node to the near position of the coordinate axis according to the modular information .

33c7dd118fa7a4d49a7253b2bbaf7f69.png

chart 3. ggClusterNet Visual layout in

ggClusterNet Functional features and future development direction

Network analysis is increasingly favored by microbial ecologists . In recent years , Developed many powerful tools , Such as Cytoscape、Gephi、igraph etc. . In terms of Visualization ,Cytoscape and Gephi It is known for its interactive graphical user interface and attractive visualization results .Cytoscape It provides powerful functions for network analysis , However, many parameters need to be adjusted to realize module visualization in the network .Gephi You can use default parameters to quickly show modules in the network , But the deep mining of the network is insufficient .ggClusterNet A combination of igraph、ggraph、sna The advantages of , Provides a variety of layout algorithms , And provide process functions (network.2() and network()) It can quickly show the modules in the network and deeply mine the network .
Future work will continue to develop optimization ggClusterNet package . In order to enhance the function of microbiome network analysis , The stability of the network will be added to the process , And deeply excavate the ecological function of the module ; Use Shiny Build a user-friendly interface , It is convenient for more researchers to conduct in-depth exploration of the network ; This will further excavate the analysis of binary networks , Develop more visual layout algorithms suitable for binary networks . All in all ,ggCLusterNet Hope to continuously improve the relevant content of network analysis , And help the majority of relevant scientific research workers .

lead   writing

Tao Wen, Penghao Xie, Shengdie Yang, Guoqing Niu, Xiaoyu Liu, Zhexu Ding, Chao Xue, Yong-Xin Liu, Qirong Shen, Jun Yuan. 2022. ggClusterNet: An R package for microbiome network analysis and modularity-based multiple network layouts. iMeta 1: e32. https://doi.org/10.1002/imt2.32

Author's brief introduction

fbb29850d0e57a60267b7fafc7d3f628.png

Wen Tao ( First author )

●   Zhongshan Young Researcher, Nanjing Agricultural University ,iMeta Journal youth Editorial Board

Focus on soil borne disease microbial process research , Good at using various biological information tools to solve ecological problems . Developed ggClusterNet, EasyStat etc. R package , EasyAmplicon, EasyMetabolome And so on . stay iMeta、ISME、Micribiome、Fundamental Research、Horticulture Research、SEL、BMC Plant Biology And other journals have published many articles .

07c1c1935931bbb2ddd11da3884fa18a.png

Yuan Jun ( Corresponding author )

●   Associate professor of Nanjing Agricultural University , Academician assistant . His research interests focus on the process of rhizosphere metabolite mediated rhizosphere interaction .

●  The main research contents include :1. The interaction between plant and microorganism during the occurrence of soil borne diseases ;2. Research on big data integration of environmental microorganisms ;3. Regulation mechanism and technology development of rhizosphere microecology . As the first author or corresponding author in ISME J,Microbiome,Fundamental Research,iMeta, PCE,SBB,Horticulture Research,SEL,BMC plant biology And other journals have published more than 20 articles , Be cited beyond 1500 Time .

More recommendations

(▼ Click the jump )

iMeta Chinese translation of articles + Video interpretation

iMeta cover | One stop tool set for macro proteomics analysis iMetaLab Suite( University of Ottawa, Canada Figeys Group )

9bedfa286fe78d441d61eafe899431c7.png

▸▸▸▸

iMeta | Wang Zhang group of South China Normal University reviewed the secret relationship between human lung microbiome and human health and disease

5879166738214dd99049f1019c94e985.png

▸▸▸▸

iMeta | Nanopore sequencing of Xiayu formation of South University of science and technology reveals that microorganisms can reduce greenhouse gas emissions from high-altitude frozen soil

19a73074a270ec8f8ba581bd57edea1d.png

▸▸▸▸

iMeta | Chen Feng of Peking University / Chenzhibin et al. Published the experimental method of sampling each part in the study of oral microbiome (Protocol)

c60022388988fb8450dea3bcbc3db64d.png

▸▸▸▸

iMeta | Zeng Zhenling of South China Agricultural University / Xiongwenguang et al - Study on the coexistence of drug resistance genes between pet dogs and their owners in the family

395c61a26f62140765f15f7224eb7d68.png

▸▸▸▸

iMeta | Mayingfei group of Shenzhen Advanced Research Institute has developed a method for analyzing intestinal flora based on neural network

821aa8f215db087c1be13645c4fa1fbc.png

▸▸▸▸

iMeta | Chenlianmin, et al. Of Nanjing Medical University summarized the potential impact of Enterobacteriaceae on complex diseases from the perspective of genome function

b0f77a4e6a33b591df8f7f0854b2075c.png

▸▸▸▸

iMeta | A review of CHENFENG group in Peking University standardization of oral microbiome : From technology driven to hypothesis driven

81b6566ac2ff9cf746bc3fcba40181a7.png

▸▸▸▸

iMeta | Linhao group of University of Electronic Science and technology of China has developed a tool for predicting protein lysine lactation sites DeepKla

87153f8914b3a336043d97a5a4a272dd.png

▸▸▸▸

iMeta | Chinese Academy of Sciences Li Xiaofang et al. Dietary licorice promotes cadmium detoxification in mice and regulates intestinal flora metabolism

3a1cf1b455d5a02283ff092d2309c0b9.png

▸▸▸▸

iMeta | Ni Yan formation of Zhejiang University MetOrigin Trace the source of modern metabolites and analyze the integration of intestinal microbiome and metabolome

418e658777179df7ecf0b25f99108f46.png

Journal Introduction

9d702b5a88ad6eaaef858ff5510e19ce.png

“iMeta” By Willie 、 Open access journal published by Enterobacteriaceae branch in cooperation with hundreds of Chinese scientists in this field , The editor in chief is Liu Shuangjiang, a researcher of the Institute of Microbiology of the Chinese Academy of Sciences, and Fu Jingyuan, a professor of Groningen University in the Netherlands . The purpose is to publish original research 、 Methods and reviews to promote metagenomics 、 Development of microbiome and bioinformatics . The goal is to publish 10%(IF > 15) High impact papers . Journal features include video contributions 、 Repeatable analysis 、 Picture polishing 、 Young editorial board members 、 front 3 Annual free publishing fee 、50 Social media publicity of 10000 users, etc .2022 year 2 It was officially launched in January !

Contact us

iMeta Home page :http://www.imeta.science

Press. :https://onlinelibrary.wiley.com/journal/2770596x
contribute :https://mc.manuscriptcentral.com/imeta
mailbox :office@imeta.science

  WeChat official account  

iMeta

  editor  

Slightly  


thank
Similar articles

2022-06-23