Application of fuzzy logic principles to the classification of 2D-PAGE maps belonging to human pancreatic cancers treated with Trichostatin-A

2D gel-electrophoresis is the most widespread technique applied in the study of cellular protein patterns. The complexity of this technique is reflected in the production of complex samples (2D gel maps) which are very difficult to compare. A method is proposed based on fuzzy logic principles. The method can be summarised in four steps: 1) digitalisation of the image; 2) fuzzyfication of the digitalised image in order to consider the natural variability of the applied technique; 3) comparison of the samples and calculation of a similarity matrix; 4) application of multidimensional scaling technique to the similarity matrix, in order to verify the presence of clusters of samples. The investigated dataset consists of IS samples belonging to two different cell lines (Paca 44 and T3M4) of control (untreated) and drug-treated pancreatic ductal carcinoma cells.