GeneMANIA: a real-time multiple association network integration algorithm for predicting gene function

By Sara Mostafavi, Debajyoti Ray, David Warde-Farley, Chris Grouios, Quaid D. Morris. 2008

1,058 citations recorded by the source corpus.

Background:Most successful computational approaches for protein function prediction integrate multiple genomics and proteomics data sources to make inferences about the function of unknown proteins. The most accurate of these algorithms have long running times, making them unsuitable for real-time protein function prediction in large genomes. As a result, the predictions of these algorithms are stored in static databases that can easily become outdated. We propose a new algorithm, GeneMANIA, that is as accurate as the leading methods, while capable of predicting protein function in real-time.Results:We use a fast heuristic algorithm, derived from ridge regression, to integrate multiple functional association networks and predict gene function from a single process-specific network using label propagation. Our algorithm is efficient enough to be deployed on a modern webserver and is as accurate as, or more so than, the leading methods on the MouseFunc I benchmark and a new yeast function prediction benchmark; it is robust to redundant and irrelevant data and requires, on average, less than ten seconds of computation time on tasks from these benchmarks.Conclusion:GeneMANIA is fast enough to predict gene function on-the-fly while achieving state-of-the-art accuracy. A prototype version of a GeneMANIA-based webserver is available at http://morrislab.med.utoronto.ca/prototype.

Subfields: Genetics, Molecular Biology, Graph Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

In Graph Learning

Citations in the corpus

Predicting multicellular function through multi-layer tissue networks (2017, background): “Notice that we exclude GeneMania in the comparison because it is not amenable to transfer learning.”
GeneMANIA Cytoscape plugin: fast gene function predictions on the desktop (2010, positive): “The GeneMANIA Cytoscape plugin integrates association networks from multiple sources into a single composite network using a conjugate gradient optimization algorithm as described in Mostafavi et al. (2008).”
Multi-view prediction of protein function (2011, neutral): “Approaches include kernel methods [15, 20] or label propagation on a network whose nodes are proteins and edges indicate similarity according to some data source [9, 19, 27].”
GeneMANIA Prediction Server 2013 Update (2013, methodology): “The transformation of mRNA expression proﬁle from GEO to a co-expression network is described later in this article, and the transformation of protein domain proﬁles to ‘shared domain’ networks has been previously described (2).”
GeneMANIA update 2018 (2018, methodology): “Technical details on the label propagation algorithms used by GeneMANIA to find related genes (3–5) and the algorithm used to assign weights to networks (3,6) have been published separately.”
Using the Gene Ontology Hierarchy when Predicting Gene Function (2009, methodology): “We have previously shown (Mostafavi et al., 2008) in unbalanced problems, we can achieve a large gain in classification performance by setting the label bias k of unlabeled nodes to be the k = n +−n− n++n− , the mean of the labels of the labelled nodes.”
A Survey of Statistical Network Models (2009, background): “Label propagation (node similarity) in networks can be harnessed to help with functional gene annotation [226].”
Genome-wide analysis identifies 12 loci influencing human reproductive behavior (2016, neutral): “We applied the GeneMANIA algorithm together with its large set of accompanying functional association data 87 .”
Selecting causal genes from genome-wide association studies via functionally-coherent subnetworks (2014, background): “Both strategies have been demonstrated to provide high-quality gene-function predictions for (amongst others) H. sapiens 19,24 , M. musculus 17,48,49 , D. rerio 50 and S. cerevisiae 51 .”
Network propagation: a universal amplifier of genetic associations (2017, methodology): “Finally, GeneMANIA 50 used a two-stage prediction process, in which multiple sources of data are first combined into a single function-specific association network, and then a propagation-based approach is applied to the resulting network to predict gene function.”
Ieee/acm Transactions on Computational Biology and Bioinformatics (neutral): “Most of these approaches predict protein functions from protein interaction networks [1] [2] [13] [17] [28] , or protein structure [14, 25].”
deepNF: deep network fusion for protein function prediction (2017, methodology): “For example, GeneMANIA [22, 20] is a widely used method that first integrates kernels of different network types into a single kernel by solving constrained linear regression problem; then, it applies Gaussian label propagation on the result-”
Understanding multicellular function and disease with human tissue-specific networks (2015, positive): “We collected tissue-specific functional interactions for each tissue from known functional relationships and low-throughput tissue-gene expression data, and mapped tissue-gene annotations from the Human Protein Reference Database 20 (HPRD) to the BRENDA Tissue Ontology 21 (BTO).”