The dplyr package recently defined a new `exprs` function that masks the `Biobase` one, messing up the normalization step in the SINCERA pipeline. The workaround is to manually specify `exprs=Biobase::exprs ` in the code, but including the specification in the package may save some frustrating debugging to users :) Issue with workaround: https://support.bioconductor.org/p/109128/