Text this: Estimating the Proportion of True Null Hypotheses in Multiple Testing Problems