how much non-thesis work should a GRA do?

I am a PhD student, and am being asked to do increasing amounts of non-thesis work. Some of this is general lab paperwork and small extra projects, which I'm happy to do and expect to lead to opportunities to present or publications. However, one project is run on clinical samples as patients present to the hospital, so is very unpredictable and time consuming (about a day's work per sample, split over 2 days). When originally asked to take part, I was told I would have to run samples only when my PI or the main student on the project was unavailable. I am now running almost all the samples, and frequently being phoned or emailed late at night to help with the project, as well as being told to come in to run samples at weekends. This project will be presented by a resident who recruits cases (but I actually collect and process most of the samples) and it is unclear if I will be on the paper. How can i gently point out to my PI that this is interfering with my thesis work and my family life, without risking endangering her future reference?