# How can limma be used to determine genes predictive of a given phenotype?

Using limma within Bioconductor here, which is a linear model package for microarray data and lmFit on gene expression data, I can perform a series of regressions gene ~ covariates by specifying design = model.matrix(~ covariates) and calling lmFit(eset, design).

How can I instead run the regressions phenotype ~ gene + covariates?

• I think you can ask this question better at bioc support site. There the developers of limma can answer you.
– benn
Apr 17 '19 at 7:49