The R package netgwas provides a set of tools based
on undirected graphical models for accomplishing three important
and interrelated goals in genetics: (1) linkage map construction,
(2) reconstructing intra- and inter-chromosomal conditional
interactions (linkage disequilibrium) network, and (3) exploring
high-dimensional genotype-phenotype network and genotype-phenotype-
environment interactions network. The netgwas can deal with biparental
species with any ploidy level.
The package implemented the recent improvements both for construction
of linkage maps in diploid and polyploid species in Behrouzi and Wit(2017b),
and in reconstructing networks for non-Gaussian data, ordinal data, and
mixed continuous and discrete data in in Behrouzi and Wit (2017a). One
application is to uncover epistatic interactions network, where the network
captures the conditionally dependent short- and long-range linkage disequilibrium
structure of a genomes and reveals aberrant marker-marker associations that are
due to epistatic selection rather than gametic linkage.
In addition, Behrouzi and Wit(2017c) implemented their proposed method to explore
genotype-phenotype network where nodes are either phenotypes or genotypes, and each
phenotype is connected by an edge to a genotype if there is a direct association
between them, given the rest of the variables. Different phenotypes may also
interconnect. Here, the conditionally dependent relationships between markers on
a genome and phenotypes is determined through Gaussian copula graphical model.
We remark that environment variables can also be included along with genotype-phenotype
input data to reconstruct networks between genotypes, phenotypes, and environment
variables. Beside, the package contains functions for simulation and visualization,
as well as three multivariate datasets taken from literature.
1. Behrouzi, P., and Wit, E. C. (2017a). Detecting Epistatic Selection with Partially Observed Genotype Data Using Copula Graphical Models. arXiv preprint, arXiv:1710.00894. 2. Behrouzi, P., and Wit, E. C. (2017b). De novo construction of q-ploid linkage maps using discrete graphical models. arXiv preprint, arXiv:1710.01063. 3. Behrouzi, P., and Wit, E. C. (2017c). netgwas: An R Package for Network-Based Genome-Wide Association Studies. arXiv preprint, arXiv:1710.01236.