Resources
The source codes for network analysis using R+igraph are available. The tutorial mentions centrality analysis, community detection, network controllability, etc. R+igraphでネットワーク分析するための様々なサンプルコードが利用可能です。中心性解析、コミュニティ検出、ネットワーク可制御性などを取り上げています。
ECOSMOS is an Estimator of COmmunity Structure based on MetabOlic networkS. ECOSMOS reveals microbial community structure; in particular, it estimates cooperative and competitive relationships based on metabolic interactions between microbes, from metagenomic data.
A single perturbation known as the universal adversarial perturbation (UAP) can foil most classification tasks conducted by deep natural networks (DNNs). This is a simple iterative method for generating a UAP that causes the DNN to classify most input images into a specific class (Hirano and Takemoto, 2020). Our method is also available in the Adversarial Robustness Toolbox (ver. 1.4 and later), a Python library for machine learning security.
MIPMET is a database on MIcrobial Physiology and METabolism. The data is useful for studies on microbe–environment interactions from a metabolic network perspective (and also genomic perspective, of course). Currently, the data on physiology and metabolism for 1,859 organisms are available.
Metabolic network analysis strongly depends on the accuracy of metabolic information (e.g., Takemoto, 2011, Takemoto, 2014). I constructed the new dataset on metabolic networks (Takemoto, 2014); but, the networks are undirected. The directed version of the networks is downloadable from here.
The methods, proposed by Feng and Takemoto (2014), estimate the largest eigenvalue of weighted bipartite networks, They are useful for evaluating a mathematical relationship between networks and stability. The source codes and 40 empirical mutualistic networks are available.