Abstract
A number of methods have been established to estimate the proportion true null hypotheses in multiple testing under the assumption of independency. On other hand, the test statistics are either discrete or continuous. In this paper, we will review an existing likelihood approach for estimating the proportion of true nulls in controlling the false discovery rates when the test statistics are continuous (see Hualing and Hanfeng, 2021). We therefore present an extension of these method that can successfully make some improvement of the performance. Simulation study demonstrates that the new estimator performs very well.