A comparison of two classes of methods for estimating false discovery rates in microarray studies.

The goal of many microarray studies is to identify genes that are differentially expressed between two classes or populations. Many data analysts choose to estimate the false discovery rate (FDR) associated with the list of genes declared differentially expressed. Estimating an FDR largely reduces to estimating π 1, the proportion of differentially expressed genes among all analyzed genes. Estimating π 1 is usually done through P-values, but computing P-values can be viewed as a nuisance and potentially problematic step. We evaluated methods for estimating π 1 directly from test statistics, circumventing the need to compute P-values. We adapted existing methodology for estimating π 1 from t- and z-statistics so that π 1 could be estimated from other statistics. We compared the quality of these estimates to estimates generated by two established methods for estimating π 1 from P-values. Overall, methods varied widely in bias and variability. The least biased and least variable estimates of π 1, the proportion of differentially expressed genes, were produced by applying the "convest" mixture model method to P-values computed from a pooled permutation null distribution. Estimates computed directly from test statistics rather than P-values did not reliably perform well.