Selects one model with variable selection FDR controlled at a specified level. A q-value for each potential variable is also returned. The input, variable selection counts over many bootstraps for several levels of penalization, is modeled as coming from a beta-binomial mixture distribution.
