Immunovia applies front-edge bioinformatic workflows and tools for analyzing microarray data, including data handling, normalization, biomarker signature identification and classification. The microarray data generated is used to build Immunovia’s own predictive model based on biomarker signatures and the support vector machine (SVM) classification algorithm.
Biomarker signatures able to accurately classify/distinguish a disease state are identified through a stepwise biomarker/analyte selection process. One biomarker is excluded at a time in an iteration process adopting Immunovia’s in-house developed backward elimination approach. The backward elimination algorithm, based on SVM classification, generates the combination of biomarkers showing the highest classification power between healthy and disease states.