Ordering of Omics Features Using Beta Distributions on Montecarlo p-Values

The current trend in genetic research is the study of omics data as a whole, either combining studies or techniques. This raises need for new robust statistical methods that can integrate and order relevant biological information. A good way to approach problem features studied according different kinds so key point associate values permit us sorting them. These are usually p-values corresponding hypothesis which has been tested each feature studied. Montecarlo method certainly one most testing. However, large number simulations needed obtain reliable p-value, becomes computationally infeasible many situations. We propose genes their differential by using score defined from beta distribution fitted generated p-values. Our simulated colorectal cancer datasets Infinium methylationEPIC array, Affymetrix gene expression array Illumina RNA-seq platforms. results show this allows proper ordering much lower than with method. Furthermore, be interpreted an estimated p-value compared other approaches like moderated t-tests. have also identified pattern eighteen common all microarrays, i.e., 21 datasets. Thus, proposed effective obtaining shows slightly smaller type I error small sizes p-value. II obtained but these differences highly reduced larger sample higher false discovery rates. Similar performances errors enable clear being evaluated.