We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
COMPUTATIONAL MODELLING OF THE 5a -REDUCTASE INHIBITORS BASED ON THE MIA-SAR APPROACH AND DESIGN OF NEW COMPOUNDS.
- Authors
AMRI, FAEZEH; NIAZI, ALI; YAZDANIPOUR, ATISA; MOMENI-ISFAHANI, TAHEREH
- Abstract
5a -reductase 2 is an interesting pharmaceutical target for the treatment of several diseases, including prostate cancer, benign prostatic hyperplasia, male pattern baldness and hirsutism. Quantitative structure-activity relationship (QSAR) analysis has been carried out for the prediction of inhibitory activity of a set of 4-X-17-Y-4-Azaandrost-3-ones as 5a -reductase 2 inhibitors. Bi-dimensional images were applied to calculate some pixels and partial least squares (PLS) algorithm was applied to QSAR modelling of 5a -reductase inhibitors. In this paper, we surveyed the effect of variable selection by application of genetic algorithms (GAs) for the PLS model. The GAs is very helpful in the variable selection in modeling and selecting the subset of pixels with the low prediction error. Pre-processing methods such as orthogonal pixel correction (OPC) were also used to provide the suitable input for modeling. These models were applied to the prediction of the molecules inhibition, which were not in the modeling procedure. The resulted model showed a high predictive ability with the root mean square error of prediction (RMSEP) of 0.52, 0.35 and 0.94 for PLS, GA-PLS and OPC-GA-PLS models respectively. Furthermore, the proposed QSAR model with the OPC-GA-PLS method was developed to predict the inhibitory activity of the new compounds.
- Subjects
BALDNESS; STANDARD deviations; BENIGN prostatic hyperplasia; STRUCTURE-activity relationships
- Publication
Journal of Science & Arts, 2019, Vol 19, Issue 4, p1011
- ISSN
1844-9581
- Publication type
Article