Vine copula graphical models in the construction of biological networks

Date

2021-01-01

Author

Farnoudkia, Hajar
Purutçuoğlu Gazi, Vilda

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

161
views

0
downloads

The copula Gaussian graphical model (CGGM) is one of the major mathematical models for high dimensional biological networks which provides a graphical representation, espe-cially, for sparse networks. Basically, this model uses a regression of the Gaussian graphical model (GGM) whose precision matrix describes the conditional dependence between the variables to estimate the coefficients of the linear regression model. The Bayesian inference for the model parameters is used to overcome the dimensional limitation of GGM under sparse networks and small sample sizes. But from the application in bench-mark data sets, it is seen that although CGGM is successful in certain systems, it may not fit well for non-normal multivariate observations. In this study, we propose the vine copulas to relax the strict normality assumption of CGGM and to describe networks from a variety of copulas alternates besides the Gaussian copula. Accordingly, we evaluate the best fitted bivariate copula distribution for every pairwise gene and compute the estimated adjacency matrix which denotes the presence of an edge between the corresponding genes. We assess the performance of our proposed approach in three network data via distinct accuracy measures by comparing the outputs with the results of the CGGM.

URI

https://hdl.handle.net/11511/91975

Journal

HACETTEPE JOURNAL OF MATHEMATICS AND STATISTICS

DOI

https://doi.org/10.15672/hujms.728352

Collections

Department of Statistics, Article

Suggestions

OpenMETU
Core

Loop-based conic multivariate adaptive regression splines is a novel method for advanced construction of complex biological networks Ayyıldız Demirci, Ezgi; Purutçuoğlu Gazi, Vilda; Weber, Gerhard Wilhelm (2018-11-01) The Gaussian Graphical Model (GGM) and its Bayesian alternative, called, the Gaussian copula graphical model (GCGM) are two widely used approaches to construct the undirected networks of biological systems. They define the interactions between species by using the conditional dependencies of the multivariate normality assumption. However, when the system's dimension is high, the performance of the model becomes computationally demanding, and, particularly, the accuracy of GGM decreases when the observations...
Bernstein approximations in glasso-based estimation of biological networks Purutçuoğlu Gazi, Vilda; Wit, Ernst (2017-03-01) The Gaussian graphical model (GGM) is one of the common dynamic modelling approaches in the construction of gene networks. In inference of this modelling the interaction between genes can be detected mainly via graphical lasso (glasso) or coordinate descent-based approaches. Although these methods are successful in moderate networks, their performances in accuracy decrease when the system becomes sparser. We here implement a particular type of polynomial transformations, called the Bernstein polynomials, of...
MARS as an alternative approach of Gaussian graphical model for biochemical networks AYYILDIZ DEMİRCİ, EZGİ; Agraz, Melih; Purutçuoğlu Gazi, Vilda (Informa UK Limited, 2017-01-01) The Gaussian graphical model (GGM) is one of the well-known modelling approaches to describe biological networks under the steady-state condition via the precision matrix of data. In literature there are different methods to infer model parameters based on GGM. The neighbourhood selection with the lasso regression and the graphical lasso method are the most common techniques among these alternative estimation methods. But they can be computationally demanding when the system's dimension increases. Here, we ...
TRACEMIN Fiedler A Parallel Algorithm for Computing the Fiedler Vector Manguoğlu, Murat; Saied, Faisal; Sameh, Ahmed (null; 2010-06-25) The eigenvector corresponding to the second smallest eigenvalue of the Laplacian of a graph, known as the Fiedler vector, has a number of applications in areas that include matrix reordering, graph partitioning, protein analysis, data mining, machine learning, and web search. The computation of the Fiedler vector has been regarded as an expensive process as it involves solving a large eigenvalue problem. We present a novel and efficient parallel algorithm for computing the Fiedler vector of large graphs bas...
Novel model selection criteria on sparse biological networks Bulbul, G. B.; Purutçuoğlu Gazi, Vilda; Purutcuoglu, E. (Springer Science and Business Media LLC, 2019-09-01) In statistical literature, gene networks are represented by graphical models, known by their sparsity in high dimensions. In this study, we suggest novel model selection criteria, namely, ICOMP, CAIC and CAICF to apply on simulated gene networks when selecting an optimal model among alternative estimated networks' constructions. In this description, we build models with the Gaussian graphical model (GGM) and the inference of GGM is achieved via the graphical lasso method. In the assessment of our proposed m...

Citation Formats

H. Farnoudkia and V. Purutçuoğlu Gazi, “Vine copula graphical models in the construction of biological networks,” HACETTEPE JOURNAL OF MATHEMATICS AND STATISTICS, vol. 50, no. 4, pp. 1172–1184, 2021, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/91975.