New paper about Dynamic Fuzzy Modeling accepted on Bioinformatics

Marco S. Nobile, Associate Professor with the Ca' Foscari University of Venice, Italy

New paper about Dynamic Fuzzy Modeling accepted on Bioinformatics

Many years ago we started a cooperation with the Department of Biotechnology and Biosciences of the University of Milano-Bicocca. The goal was to govern programmed cell death on K-ras cancer cells under conditions of glucose starvation.

Glucose is directly related to metabolism, i.e., glycolysis and TCA cycle in mitochondria, which in turn determine the production of ATP. ATP regulates several processes: autophagy, necrosis, apoptosis (mediated by Caspase proteases) and of course promotes cell survival. Glucose is also involved in a side pathway of glycolysis, that is, the hexosamine biosynthetic pathway, which provides material for N-glycosylation. The latter process determines the attachment of the cell and the activation of Unfolded Protein Response (UPR), who affects calcium and (indirectly) both necrosis and apoptosis by the intermediate activity of Bcl-2. Our model also considered the effect of ROS (controlled by mitochondria and affecting Caspase 3, autophagy, and necrosis) and was also able to describe low and high activation of protein kinase A.

The development of a single model encompassing all these molecules, pathways, and processes was more than challenging. A detailed mechanistic representation of all biochemical processes was clearly unfeasble. In order to handle the inherent heterogeneity of this complex system, while preserving the emergent dynamic behavior, we created a novel modeling approach named Dynamic Fuzzy Modeling (DFM).

Fuzzy Logic allowed us to represent all interactions in terms of human understandable logical statements. By using glucose as input variable and by repeating the fuzzy inference, we managed to produce a simulated trajectory that fit the experimental data. Moreover, we exploited Simulated Annealing (SA) to identify novel promising perturbations of the system able to maximize apoptosis and minimize survival. For instance, our method determined that the activation of UPR, by means of thapsigargin, coupled to autophagy inhibition, using chloroquine, can yield an increase of cell death especially under glucose starvation. Our laboratory experiments proved that the prediction of the DFM, guided by SA, was actually correct.

The paper was accepted for publication on Bioinformatics.

No Comments

Add your comment