Using PPI network autocorrelation in hierarchical multi-label classification trees for gene function prediction

Abstract

BackgroundOntologies and catalogs of gene functions, such as the Gene Ontology (GO) and MIPS-FUN, assume that functional classes are organized hierarchically, that is, general functions include more specific ones. This has recently motivated the development of several machine learning algorithms for gene function prediction that leverages on this hierarchical organization where instances may belong to multiple classes. In addition, it is possible to exploit relationships among examples, since it is plausible that related genes tend to share functional annotations. Although these relationships have been identified and extensively studied in the area of protein-protein interaction (PPI) networks, they have not received much attention in hierarchical and multi-class gene function prediction. Relations between genes introduce autocorrelation in functional annotations and violate the assumption that instances are independently and identically distributed (i.i.d.), which underlines most machine learning algorithms. Although the explicit consideration of these relations brings additional complexity to the learning process, we expect substantial benefits in predictive accuracy of learned classifiers.ResultsThis article demonstrates the benefits (in terms of predictive accuracy) of considering autocorrelation in multi-class gene function prediction. We develop a tree-based algorithm for considering network autocorrelation in the setting of Hierarchical Multi-label Classification (HMC). We empirically evaluate the proposed algorithm, called NHMC (Network Hierarchical Multi-label Classification), on 12 yeast datasets using each of the MIPS-FUN and GO annotation schemes and exploiting 2 different PPI networks. The results clearly show that taking autocorrelation into account improves the predictive performance of the learned models for predicting gene function.ConclusionsOur newly developed method for HMC takes into account network information in the learning phase: When used for gene function prediction in the context of PPI networks, the explicit consideration of network autocorrelation increases the predictive performance of the learned models. Overall, we found that this holds for different gene features/ descriptions, functional annotation schemes, and PPI networks: Best results are achieved when the PPI network is dense and contains a large proportion of function-relevant interactions.

ComiRNet: a web-based system for the analysis of miRNA-gene regulatory networks

“…This is in line with recent research on gene function classification when PPI network data can be exploited [45]. …”

Machine learning for discovering missing or wrong protein function annotations

“…As its name suggests, the algorithm adaptation strategy consists of adapting a traditional algorithm to handle hierarchical constraints. Masera and Blanzieri [6] created a neural network whose architecture incorporates the underlying hierarchy, making gradient updates flow from the neurons associated to the leaves up neurons associated to their parent nodes; Sun et al [8] proposed to use Partial Least Squares to reduce both label and feature dimension, followed by an optimal path selection algorithm; Barros et al [17] proposed a centroid based method where the training data is initially clustered, then predictions are performed by measuring the distance between the new instance and all clusters, the label set associated to the closest cluster is given as the prediction; Borges and Nievola [31] developed a competitive neural network whose architecture replicates the hierarchy; Vens et al [2] also proposed to train a single Predictive Clustering Tree for the entire hierarchy; as an extension of [2], Schietgat et al [21] proposed to use ensemble of Predictive Clustering Trees; Stojanova et al [18] proposed a slight modification for Predictive Clustering Trees in which the correlation between the proteins is also used to build the tree.…”

Modelling Progressive Filtering

“…In particular, let us recall the works concerning decision trees (e.g., [45], [30], and [40]), neural networks (e.g., [37] and [9]), support vector machines (e.g., [43], [12] and [6], and evolutionary computation (e.g., [23], [24], [36] and [10]). …”