AptRank: an adaptive PageRank model for protein function prediction on   bi-relational graphs

Abstract

. Motivation. Diffusion-based network models are widely used for protein function prediction using protein network data and have been shown to outperform neighborhood-based and module-based methods. Recent studies have shown that integrating the hierarchical structure of the Gene Ontology (GO) data dramatically improves prediction accuracy. However, previous methods usually either used the GO hierarchy to refine the prediction results of multiple classifiers, or flattened the hierarchy into a function-function similarity kernel. No study has taken the GO hierarchy into account together with the protein network as a two-layer network model.Results. We first construct a Bi-relational graph (Birg) model comprised of both protein-protein association and function-function hierarchical networks. We then propose two diffusion-based methods, BirgRank and AptRank, both of which use PageRank to diffuse information on this two-layer graph model. BirgRank is a direct application of traditional PageRank with fixed decay parameters. In contrast, AptRank utilizes an adaptive diffusion mechanism to improve the performance of BirgRank. We evaluate the ability of both methods to predict protein function on yeast, fly, and human protein datasets, and compare with four previous methods: GeneMANIA, TMC, ProteinRank and clusDCA. We design three different validation strategies: missing function prediction, de novo function prediction, and guided function prediction to comprehensively evaluate predictability of all six methods. We find that both BirgRank and AptRank outperform the previous methods, especially in missing function prediction when using only 10% of the data for training.Conclusion. AptRank naturally combines protein-protein associations and the GO function-function hierarchy into a two-layer network model without flattening the hierarchy into a similarity kernel. Introducing an adaptive mechanism to the traditional, fixed-parameter model of PageRank greatly improves the accuracy of protein function prediction.

Benchmarking network propagation methods for disease gene identification

“…Diffusion (propagation) methods are central in this study. We used the random walk-based personalised PageRank [25], previously used in similar tasks [26], as implemented in igraph [27]. The remaining diffusion-based methods were run on top of the regularised Laplacian kernel [28], computed through diffuStats [29].…”

Adaptive Diffusions for Scalable Learning Over Graphs

“…In this sense, AdaDIF undertakes statistical learning in the space of functional rankings, tailored to the underlying semi-supervised classification task. A related method termed AptRank was recently proposed in [46] specifically for protein function prediction. Differently from AdaDIF the method exploits meta-information regarding the hierarchical organization of functional roles of proteins and it performs random walks on the heterogeneous protein-function network.…”

“…Celebrated representatives include those based on the Personalized PageRank (PPR) and the Heat Kernel that were found to perform remarkably well in certain application domains [22], and have been nicely linked to particular network models [23], [3], [24]. Spectral diffusions have been used for community detection [47], [45], where local diffusion patterns are produced to approximate low-conductance communities, and adaptive PPR has been applied for prediction on a heterogeneous protein-function network [46].…”

Accurate and Efficient Gene Function Prediction using a Multi-Bacterial Network

“…For instance, fewer than 0.01% (about 10,000 out of 104 million) of prokaryotic genes in the UniProt Knowledgebase [2] have a Gene Ontology (GO) [3] annotation with an experimental evidence code. To address this gap, several computational methods have been developed to associate molecular functions and biological processes with genes lacking experimental annotations [4][5][6][7][8][9][10][11]. These predicted associations assist in prioritizing follow-up experiments to determine the biological functions performed by gene products [12].…”

“…These predicted associations assist in prioritizing follow-up experiments to determine the biological functions performed by gene products [12]. Network-based techniques are among the most accurate and widely-used approaches for gene function prediction [4][5][6]9].…”