MicroGenix integrates multi-omics data to identify the genotype-microbiome interactions in shaping molecular phenotype.
The MicroGenix R package is easily installed from the GitHub repository:
install.packages("devtools")
devtools::install_github("MicrobeLab/MicroGenix")
The MicroGenix R package includes three main functions. Detailed documentation is available through the R help interface (?MicroGenixTrain, ?MicroGenixPredict, ?MicroGenixAssociation). Demo input files for running examples below are available in the extdata folder.
MicroGenixTrain is used to fit models to predict gene expression with elastic net. Predictors include SNPs, microbes, and SNP-microbe interactions.
input_taxa <- 'extdata/example_taxon_abundance.csv'
input_geno <- 'extdata/example_genotype_dosage.csv'
input_expr <- 'extdata/example_gene_expression.csv'
fit_data <- MicroGenixTrain(input_taxa, input_geno, input_expr, output_prefix = 'example_output_model')
input_taxa and input_geno are comma-separated table files in shape [number_of_samples, number_of_taxa] and [number_of_samples, number_of_SNPs], respectively. input_expr is a file in shape [number_of_samples,] with the expression levels of a single gene.
MicroGenixPredict predicts gene expression using models trained with MicroGenixTrain.
input_taxa <- 'extdata/example_taxon_abundance.csv'
input_geno <- 'extdata/example_genotype_dosage.csv'
model <- 'example_output_model.rds'
predicted_data <- MicroGenixPredict(model, input_taxa, input_geno)
input_taxa and input_geno are in the same format as the input files for MicroGenixTrain. model is a model file generated using MicroGenixTrain.
MicroGenixAssociation performs association tests between phenotype and gene expression levels predicted using MicroGenixPredict.
pred_expr <- predicted_data$predicted_expr
metadata <- 'extdata/example_metadata.csv'
assoc_result <- MicroGenixAssociation(pred_expr, metadata, pheno = 'pheno')
metadata is a comma-separated file with phenotype and covariates.
Bugs and difficulties in using MicroGenix are welcome on the issue tracker.
The MicroGenix method is described in "Multi-omics integration unravels genotype-microbiome interactions shaping the conjunctival transcriptome". Models trained using the conjunctival multi-omics data are available here.