Read papers ,you need read all the papers related to your subject. then i am sure you can make predictions for most of your experiments you can gonna conduct and it would be a problem to write some abstract even you don't do any experiments. if you cannot, it means you don't read enough papers. in other words, you should go on reading.