Recent advances in predicting gene–disease associations

Opap, Kenneth; Mulder, Nicola

doi:10.12688/f1000research.10788.1

Cited by 38 publications

(23 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result various computational approaches have been developed to aid the discovery of such associations, such as knowledge-based methods (e.g., (Zhou and Skolnick 2016)) and methods based on text mining (e.g., (Kolker et al 2015)), crowdsourcing (e.g., (Loguercio et al 2013)) and networks (e.g., (Singh-Blom et al 2013;Zeng et al 2017)). Comprehensive surveys of these methods can be found in (Piro and Di Cunto 2012;Opap and Mulder 2017;Seyyedrazzagi and Navimipour 2017). Tremendous heterogeneity can be found in biological data -comprising measurements from diverse aspects of our complex biological systems -that are used to infer gene-disease associations.…”

Section: Case Study: Gene-disease Association Predictionmentioning

confidence: 99%

Deep collective matrix factorization for augmented multi-view learning

Mariappan

Rajan

2019

Mach Learn

View full text Add to dashboard Cite

Learning by integrating multiple heterogeneous data sources is a common requirement in many tasks. Collective Matrix Factorization (CMF) is a technique to learn shared latent representations from arbitrary collections of matrices. It can be used to simultaneously complete one or more matrices, for predicting the unknown entries. Classical CMF methods assume linearity in the interaction of latent factors which can be restrictive and fails to capture complex non-linear interactions. In this paper, we develop the first deep-learning based method, called dCMF, for unsupervised learning of multiple shared representations, that can model such non-linear interactions, from an arbitrary collection of matrices. We address optimization challenges that arise due to dependencies between shared representations through Multi-Task Bayesian Optimization and design an acquisition function adapted for collective learning of hyperparameters. Our experiments show that dCMF significantly outperforms previous CMF algorithms in integrating heterogeneous data for predictive modeling. Further, on two tasks -recommendation and prediction of gene-disease association -dCMF outperforms state-of-the-art matrix completion algorithms that can utilize auxilliary sources of information.

show abstract

Section: Case Study: Gene-disease Association Predictionmentioning

confidence: 99%

Deep collective matrix factorization for augmented multi-view learning

Mariappan

Rajan

2019

Mach Learn

View full text Add to dashboard Cite

show abstract

“…This method applies the method of network analysis to predict the interaction between genes and diseases . The recent advances in predicting gene-disease associations have been reviewed by Opap and Mulder (2017). An understanding of the association between genetics and disease is an important step in understanding the etiology of diseases.…”

Section: Introductionmentioning

confidence: 99%

GBDTL2E: Predicting lncRNA-EF Associations Using Diffusion and HeteSim Features Based on a Heterogeneous Network

Wang

Kuang

et al. 2020

Front. Genet.

View full text Add to dashboard Cite

Interactions between genetic factors and environmental factors (EFs) play an important role in many diseases. Many diseases result from the interaction between genetics and EFs. The long non-coding RNA (lncRNA) is an important non-coding RNA that regulates life processes. The ability to predict the associations between lncRNAs and EFs is of important practical significance. However, the recent methods for predicting lncRNA-EF associations rarely use the topological information of heterogenous biological networks or simply treat all objects as the same type without considering the different and subtle semantic meanings of various paths in the heterogeneous network. In order to address this issue, a method based on the Gradient Boosting Decision Tree (GBDT) to predict the association between lncRNAs and EFs (GBDTL2E) is proposed in this paper. The innovation of the GBDTL2E integrates the structural information and heterogenous networks, combines the Hetesim features and the diffusion features based on multi-feature fusion, and uses the machine learning algorithm GBDT to predict the association between lncRNAs and EFs based on heterogeneous networks. The experimental results demonstrate that the proposed algorithm achieves a high performance.

show abstract

“…Therefore, several computational approaches were developed to aid the prediction of novel DG candidates and prioritization of the most promising ones for experimental follow-up studies. Gene prioritization methods differ in the type and number of data sources used (such as biomedical literature, gene expression data, functional annotations, and interaction networks), the prior knowledge about the DGs, the data representation, or the prediction/prioritization functions, however, the majority uses the guilty-by-association principle identifying and prioritizing candidate genes based on their topological or functional similarity (correlated expression profiles, protein interactions or participation in the same biological processes) to known DGs [31] [32].…”

Section: Introductionmentioning

confidence: 99%

Cross Disease Network Analysis

Ramos

García-Vaquero

Gama‐Carvalho

et al. 2019

2019 IEEE 6th Portuguese Meeting on Bioengineering (ENBENG)

View full text Add to dashboard Cite

Diseases are often complex, caused by a combination of several factors including genetic, environmental and lifestyle factors. The complexity makes it more challenging to uncover the pathomechanisms underlying genotype-phenotype relationships. Cellular networks offer a simple framework to represent the highly interlinked cellular systems, by reducing cellular components, such as metabolites, proteins, DNA molecules or RNA molecules, to nodes and physical, biochemical or functional interactions to links between them. Diseases can be viewed as perturbations of these cellular networks, that lead to faulty physiological functions. Different diseases can have common deregulated molecular pathways, represented in the network as an overlap of subnetworks that are affected in each disease, particularly if they partially share phenotypes. The discovery of genes associated with multiple diseases is especially interesting because it can shed light on the molecular mechanisms implicated in the commonly affected physiological functions and provide new polyvalent therapeutic targets. This dissertation builds upon a previously developed network-based method, called double specificbetweenness (S2B) method, to prioritize nodes with a higher probability of being simultaneously associated with two phenotypically similar diseases. The method was developed to use undirected networks of physical interactions between proteins and extract a network property, a modified version of betweenness centrality, to prioritize proteins specifically connected with two different diseases. The method was tested with artificial disease network modules and applied to two fatal motor neuron diseases: Amyotrophic Lateral Sclerosis and Spinal Muscular Atrophy. The present work aims to expand the S2B method enabling the analysis of networks with directed interactions. This expansion allows the analysis of signaling and transcriptional regulatory networks, providing new regulatory information that can't be captured with protein-protein interactions, contributing to richer mechanistic hypothesis to explain the common physiological deficiencies. The new extended version of the method was tested with several types of directed artificial disease modules, proving to be able to efficiently predict the network overlap between them and offer new insights into the role of the predicted candidates in the network. The directed S2B was also applied to the same motor neuron disease pair, demonstrating its ability to retrieve novel disease genes associated with regulatory mechanisms dysregulated in motor neuron degeneration.

show abstract

Recent advances in predicting gene–disease associations

Cited by 38 publications

References 54 publications

Deep collective matrix factorization for augmented multi-view learning

Deep collective matrix factorization for augmented multi-view learning

GBDTL2E: Predicting lncRNA-EF Associations Using Diffusion and HeteSim Features Based on a Heterogeneous Network

Cross Disease Network Analysis

Contact Info

Product

Resources

About