run.feature.lm: Runs a linear model for a single given phenotypic feature
In definitelysean/sgaphd: Sean Gets A PhD: Research Convenience Functions

This function is a wrapper for limma::lmFit, which is usually used to run linear models for microarray data. It returns a dataframe with the t-statistic, p-value, and FDR adjusted p-value for each gene.

1	run.feature.lm(feature.name, pheno.data, dataset, is.numeric = TRUE)

`feature.name`	A string of the column name in the phenotype data matrix corresponding to the feature of interest
`pheno.data`	The full phenotypic data matrix (e.g. pData(expression.set))
`dataset`	the expression matrix of interest. ncol(dataset) should == nrow(pheno.data)
`is.numeric`	Whether the feature in question is numeric or not. If not numeric, it's assumed to be a factor.