miniABS

miniABS (mini Absolute Breast Cancer Subtyper) is an absolute, single-sample subtype classifier for breast cancer using Random Forest model of pairwise gene expression ratios (PGER) among 11 functional genes. With a systematic gene selection and reduction step, we aimed to minimize the size of gene set without losing a functional interpretability of the classifier. We validated the model performance using a large, heterogeneous cohort that consists of multiple public datasets across four different platforms. We anticipate that the high accuracy and reproducibility of miniABS may provide a SSP at a low cost, as well as providing a platform for cross comparison among gene expression datasets generated with different technical platforms.