The Quantitative Structure-Activity Relationship (QSAR) models are a very useful tool in the design of new chemical compounds. The QSAR methods are based on the assumption that the activity of a certain chemical compound is related to its structure. Two types of QSAR analysis are summarized in this review: Linear Regression model (LR) and Linear Discriminant Analysis model (LDA).
Data from preclinical assays of ChEMBL are obtained. This data is treated mathematically and a PTML-LDA model is obtained in which the conditions are analyzed individually. With this equation and the values of the descriptors of new molecules, the activity of the compound to be tested could be predicted.
ChEMBL-tik datu basea lortuta, perturbazio teoria (PT) eta Machine Learning (ML) teknikak erabilita PTML eredu bat eraiki da, zein sistema biomolekular konplexuetan erabili daitekeen perturbazioen efektua kuantifikatzeko. Eredu hau erabilita konposatu berri batek erakusten dituen minbiziaren aurkako parametro klinikoen (ki, LD50, etab.) balioak aurresan ditzakegu. After obtaining the database from ChEMBL we combine Perturbation Theory (PT) and Machine Learning (ML) to obtain PTML Model, which has been created to quantify the perturbations of complex bio molecular systems. The model can predict preclinical (ki, LD50, etc.) values of new anti-cancer compounds.