Calculates adjusted p-values for the null hypothesis of no replicability across studies for two study designs: (i) a primary and follow-up study, where the features in the follow-up study are selected from the primary study, as described in Bogomolov and Heller (2013) <doi:10.1080/01621459.2013.829002> and Heller, Bogomolov and Benjamini (2014) <doi:10.1073/pnas.1314814111>; (ii) two independent studies, where the features for replicability are first selected in each study separately, as described in Bogomolov and Heller (2018) <doi:10.1093/biomet/asy029>. The latter design is the one encountered in a typical meta-analysis of two studies, but the inference is for replicability rather than for identifying the features that are non-null in at least one study.
