O'Reilly logo

Statistical and Machine Learning Approaches for Network Analysis by Subhash C. Basak, Matthias Dehmer

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

1.4 Reconstruction of Biological Networks

In this section, we describe some existing approaches to reconstruct directed and undirected biological networks from gene expression data and gene sets. To reconstruct directed networks from gene expression data, we present Boolean network, probabilistic Boolean network, and Bayesian network models. We discuss cGraph, frequency method and NICO approaches for network reconstruction using gene sets (Fig 1.4). Next, we present relevance networks and graphical Gaussian models for the reconstruction of undirected biological networks from gene expression data (Fig 1.5). The review of models in case of directed and undirected networks is largely based on Refs. [6–8,17–20] and [2,3,13,32], respectively.

Figure 1.4 (a) Representation of inputs and Boolean data in the frequency method from Ref. [18]. (b) Network inference from PAK pathway [67] using NICO, in the presence of a prior known end points in each path [68]. (c) The building block of cGraph from Ref. [17].

img

Figure 1.5 Comparison of correlation-based relevance networks (a) and partial correlation based graphical Gaussian modeling (b) performed on a synthetic data set generated from multivariate normal distribution. The figures represent estimated correlations and partial correlations between every pair of genes. Light to dark colors correspond to high to low correlations and partial correlations. ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required