martini is an R package to perform GWAS experiment that considers prior biological knowledge. This knowledge is modeled as a network of SNPs, were edges represent functional relationships between them (e.g. belonging to the same gene). Then, it looks for regions of the network associated with the phenotype using SConES or SigMod.
Install martini like any other Bioconductor package:
install.packages("BiocManager")
BiocManager::install("martini")Running martini is a three step process:
library(martini)
# 1. Read GWAS data with read.pedfile (or load the example :) )
data(minigwas)
# 2. Create the SNP network: GS (structural information), GM (GS + gene
# annotation information) or GI (GM + protein-protein interaction information)
gs <- get_GS_network(minigwas)
# 3. Run SConES, finding the best parameters by cross-validation
res <- scones.cv(minigwas, gs)
# the output is an igraph subnetwork containing the selected SNPs
res
# IGRAPH d9128a0 UNW- 12 10 --
# + attr: name (v/c), chr (v/n), pos (v/n), weight (e/n)
# + edges from d9128a0 (vertex names):
# [1] 1A1--1A2 1A2--1A3 1A3--1A4 1A4--1A5 1A5--1A6 2C1--2C2 2C2--2C3 2C3--2C4 2C4--2C5 2C5--2C6Please, refer to the vignettes for more detailed usage examples. martini results can be further examined using the blur package.
A more detailed description can be found in the pre-print. If you use martini in your work, please cite us:
@article{martini2021,
title = {martini: an {R} package for genome-wide association studies using {SNP} networks},
author = {Climente-González, Héctor and Azencott, Chloé-Agathe},
url = {http://biorxiv.org/lookup/doi/10.1101/2021.01.25.428047},
journal = {bioRxiv},
month = jan,
year = {2021},
doi = {10.1101/2021.01.25.428047}
}