Fuzzy and quantum methods of information retrieval to analyse genomic data from patients at different stages of fibrosis

We have recently introduced in a novel method of unsupervised clustering, we termed semantic distillation, inspired from the quantum theory of measurement, allowing to analyse statistical data and regroup objects according to their contextual specificity. Here we introduce an improvement of the method to make it applicable even in the more difficult case where data have very small statistical variability from sample to sample. We applied our method to a dataset from DNA arrays of different anatomopathological samples concerning 14 patients suffering from fibrosis. The clusters produced by our method have been automatically indexed by the evolutionary stage of the disease and the genes regrouped therein identified as responsible of biological processes specific for every stage.