DFDR2 adjustment

[ksiegmund]

Thank you for creating this package. When I run DFDR.p.adjust() and DFDR2.p.adjust(), I get the same individual-level BH-adjusted p-values with a different family-level filter. I think the individual-level BH-adjusted p-values for DFDR2.p.adjust() should be from a single BH-adjustment of the pool of individual-level tests that make it past the filter on families.