Should I run a control for my unknown patient sample?