In silico study for discovering novel thiazole derivatives as anti-breast cancer agents (MCF-7)

Abstract

The work validated the pharmacokinetic characteristics and docking approach, and
discovered novel thiazole compounds with MCF-7 breast anti-cancer efficacy using
quantitative structure-activity relationships (QSAR) models. Using an experimental dataset of
53 thiazole derivatives with IC50 values, QSAR models were developed using multivariate
linear regression (QSARMLR) and artificial neural networks (QSARANN). Eight descriptions,
with statistical values of R2train = 0.875 and Q2LOO = 0.834, comprised the successfully
developed QSARMLR model. Based on the descriptors of the QSARMLR model, which has
statistical values of R2 = 0.918, Q2test = 0.934, and Q2CV = 0.916, the QSARANN neural network
model with the architectural network of I(8)-HL(9)-O(1) has also been constructed. Using the
AD and outliers analysis, the models were utilized to forecast 120 new design derivatives based
on the thiazole framework, and the IC50 values of 10 new thiazole derivatives were determined.
Additionally, the derivatives were assessed for resistance to estrogen-positive breast cancer by
docking them onto the Polo-like kinases (Plk1) receptor and screening them for
pharmacokinetic features in accordance with Lipinski and Ghose guidelines. Consequently, it
was discovered that the thiazole TAZ5 had promising derivative activity against the primary
MCF-7 breast cancer cell line.