This function analyzes feature sets using MAGMA or Bayesian methods for association testing. It supports joint or marginal testing, as well as Bayesian linear regression using different priors (`bayesC`, `bayesR`).
magma(
stat = NULL,
sets = NULL,
method = "magma",
type = "joint",
test = "one-sided",
pi = 0.001,
nit = 5000,
nburn = 1000
)A data frame or list with analysis results.
A numeric vector or matrix of summary statistics, where rows represent features and columns represent phenotypes.
A list of feature sets (e.g., genes, SNPs) to be analyzed.
A string specifying the method to use. Options are `"magma"`, `"blr"`, `"bayesC"`, or `"bayesR"`. Default is `"magma"`.
A string specifying the type of analysis to perform. Options are `"joint"` (default) or `"marginal"`. Only used with `method = "magma"`.
A string specifying the statistical test. Options are `"one-sided"` (default) or `"two-sided"`. Only used with `method = "magma"`.
A numeric value specifying the proportion of non-zero effects. Used for Bayesian methods. Default is `0.001`.
An integer specifying the number of iterations for Bayesian methods. Default is `5000`.
An integer specifying the number of burn-in iterations for Bayesian methods. Default is `1000`.
The function uses either the MAGMA approach for set-based testing or Bayesian linear regression to estimate effect sizes and probabilities of association for feature sets. For Bayesian methods, a spike-and-slab prior is applied.
The `stat` input must have row names corresponding to feature identifiers. The `sets` input must be a named list, where each element corresponds to a feature set.