Table 9 Quantile from which the quantiles for one method are always greater than those of the other method, regarding the distributions of variable importances, across the SNPs, for T-Trees and the hybrid FLTM / T-Trees approach. For example, for (T1D, Chr13) dataset, the hybrid quantiles are always greater than the T-Trees quantiles as from top100 quantile. Moreover, the hybrid quantiles are an order of magnitude higher as from top20 quantile

From: A method combining a random forest-based technique with the modeling of linkage disequilibrium through latent variables, to run multilocus genome-wide association studies