1: \begin{abstract}

2: The proportion of false null hypotheses is a very important quantity in

3: statistical modelling and inference based on the two-component mixture model

4: and its extensions, and in control and estimation of the false discovery rate

5: and false non-discovery rate. Most existing estimators of this proportion

6: threshold p-values, deconvolve the mixture model under constraints on its

7: components, or depend heavily on the location-shift property of distributions.

8: Hence, they usually are not consistent, applicable to non-location-shift

9: distributions, or applicable to discrete statistics or p-values. To eliminate

10: these shortcomings, we construct uniformly consistent estimators of the

11: proportion as solutions to Lebesgue-Stieltjes integral equations. In

12: particular, we provide such estimators respectively for random variables whose

13: distributions have Riemann-Lebesgue type characteristic functions, form

14: discrete natural exponential families with infinite supports, and form natural

15: exponential families with separable moment sequences. We provide the speed of

16: convergence and uniform consistency class for each such estimator under

17: independence. In addition, we provide example distribution families for which

18: a consistent estimator of the proportion cannot be constructed using our

19: techniques.

20: \end{abstract}

21: