Application of genetic algorithms for pixel selection in MIA-QSAR study of CK-1δ inhibitors as neuroprotective agents

Amyotrophic Lateral Sclerosis (ALS) is a progressive neurodegenerative disease which results in paralysis, muscle wasting and death. Death of the motor neurons of the cortex, spinal cord and brain stem is a characteristic of this disease which eventually leads to death of the patient usually resulting from respiratory failure, mostly within 3-5 years from the appearance of symptoms. Among the reported isoforms and splice variants of CK-1 protein superfamily, CK-1 is known to phosphorylate different serine and threonine sites on TDP-43 protein in vitro and the thus qualifies as potential target for ALS treatment. Quantitative structure-activity relationships (QSAR) model [2] has been developed for the CK-1 inhibitors of 37 various N-Benzothiazolyl-2-Phenyl Acetamide derivatives [1]. Bidimensional images were used to calculate some pixels [2]. Multivariate image analysis [3, 4] was applied to QSAR modeling of the CK-1 inhibitors of these compounds by means of multivariate calibration such as principal component regression (PCR) and partial least squares (PLS). In this study we investigate the effect of pixel selection by application of genetic algorithms (GAs) for PLS model. GAs is very useful in the variable selection in modeling and calibration because of the strong effect of the relationship between presence/absence of variables in a calibration model and the prediction ability of the model itself. The subset of pixels, which resulted in the low prediction error, was selected by genetic algorithm. The resulted model showed high prediction ability with RMSEP of 0.0985, 0.0763 and 0.0103 for PCR, PLS and GAPLS models, respectively. Furthermore, the proposed QSAR model with GA-PLS was used for modification of structure and their activity predicted.