Clue: how is the sample coded as being seropositive or seronegative? How does this differ between the original and additional datasets?