Abstract:BackgroundRecent genomic and bioinformatic advances have motivated the development of numerous network models intending to describe graphs of biological, technological, and sociological origin. In most cases the success of a model has been evaluated by how well it reproduces a few key features of the real-world data, such as degree distributions, mean geodesic lengths, and clustering coefficients. Often pairs of models can reproduce these features with indistinguishable fidelity despite being generated by vastly different mechanisms. In such cases, these few target features are insufficient to distinguish which of the different models best describes real world networks of interest; moreover, it is not clear a priori that any of the presently-existing algorithms for network generation offers a predictive description of the networks inspiring them.ResultsWe present a method to assess systematically which of a set of proposed network generation algorithms gives the most accurate description of a given biological network. To derive discriminative classifiers, we construct a mapping from the set of all graphs to a high-dimensional (in principle infinite-dimensional) "word space". This map defines an input space for classification schemes which allow us to state unambiguously which models are most descriptive of a given network of interest. Our training sets include networks generated from 17 models either drawn from the literature or introduced in this work. We show that different duplication-mutation schemes best describe the E. coli genetic network, the S. cerevisiae protein interaction network, and the C. elegans neuronal network, out of a set of network models including a linear preferential attachment model and a small-world model.ConclusionsOur method is a first step towards systematizing network models and assessing their predictability, and we anticipate its usefulness for a number of communities.

Inferring network mechanisms: The Drosophila melanogaster protein interaction network

Discriminative topological features reveal biological network mechanisms

Inferring Interaction Networks from Multi-Omics Data.

Deducing Topology of Protein-Protein Interaction Networks from Experimentally Measured Sub-Networks

Duplication and Divergence Effect on Network Motifs in Undirected Bio-Molecular Networks

Protein Interaction Networks of Saccharomyces Cerevisiae, Caenorhabditis Elegans and Drosophila Melanogaster: Large‐scale Organization and Robustness

Graph theoretic analysis of protein interaction networks of eukaryotes

Simulated evolution of protein-protein interaction networks with realistic topology

Quantifying robustness of the gap gene network

Modular Organization of Protein Interaction Networks

Classifying Protein–Protein Binding Affinity with Free-Energy Calculations and Machine Learning Approaches

The exponential distance rule based network model predicts topology and reveals functionally relevant properties of the Drosophila projectome

Mean field theory for biology inspired duplication-divergence network model

Predicting Protein Function Based on the Topological Structure of Protein Interaction Networks

Discriminating Direct and Indirect Connectivities in Biological Networks

Evolution of complex modular biological networks

Transcription Factor Networks in Drosophila Melanogaster

Relating Three-Dimensional Structures To Protein Networks Provides Evolutionary Insights

Next-generation Large-Scale Binary Protein Interaction Network for Drosophila Melanogaster

A statistical method for revealing form-function relations in biological networks

Measuring Multiple Evolution Mechanisms of Complex Networks