Adimen Artifiziala. PTIA eredu kimioinformatikoa endekapenezko gaixotasun neurologikoen kontrako botikak aurkitzeko

Abstract

Cheminformatics is part of theoretical chemistry and consists of the use of computer techniques in pharmaceutical chemical science problems. This work presents the possible application of the IFPTIA methodology (Information Fusion + Perturbation Theory + Machine Learning) in the field of medical chemistry. Specifically, Alzheimer's, Parkinson's, Amyotrophic Lateral Sclerosis (ALS), Friedreich ataxia and Huntington degenerative neurological diseases have been studied. In the predictive model, the sequences of proteins that may be related to them and the Protein Interaction Network (PES) of the different regions of the brain (entorrinal cortex (CR), hippocampus (HIP), central temple curve (CEP), rear cortex (CSC), upper forehead curve (BB) and visual cortex (CB) have been considered. The statistical parameters of the model obtained have been good; Sn (%) = 77.76, Sp (%) = 72.69 and Ac (%) = 73.83 for training; and Sp (%) = 72.66, Sn (%) = 77.95 and Ac (%) = 73.84 for validation. These diseases are incurable and from the moment the symptoms appear, the patients are weakened until the neurons in the brain die. Among the usual problems that this entails are problems of poor movement and brain functioning. In this sense, cheminformatic models can be a very useful technique in the development of new drugs, in efforts to reduce animals for testing and, in general, to avoid resources. They can predict the probability that a compound may or may not be active in degenerative neurological diseases. These models have therefore been shown to be very useful tools for understanding the mechanisms behind these diseases, and this has led to the opening of promising therapeutic pathways.