Therefore, our new format defines several data representations, one of which is equivalent to PLINK 1 binary, and allows different files, or even variants within a single file, to use different representations. To work with this, PLINK 2.0 will include a translation layer which allows individual functions to assume a specific representation is used. As with the rest of PLINK’s source code, this translation layer will be GPLv3-licensed open source; and unlike most of the other source code, we are explicitly designing it to be usable as a standalone library. PLINK 2.0 will also be able to convert files/variants from one data representation to another, making it practical for third-party tools lacking access to the library to demand a specific representation.