Qauntitative structural activity relationship (QSAR) is a Computational modelling of predicting or generating compound with high biological activities. One of the major challenges in pharmaceutical company is to design a drug that will cure certain disease without being resistance to the disease later on. In a bid to cater for this, many drugs have been developed. QSAR is a powerful, easy, fast and effective method of deveping a new drug unlike the experimental method which is time consuming, costly and stressful. QSAR is a method of relating the physicochemical properties of a compound with it's biological activities.
For example, In a situation where scientist want to test the biological activities of more than 1000 compounds, testing of all the Compounds one by one will be stressful, costly and will waste alot of time. Instead of testing all the Compounds, Only 50 out of 1000 compounds can be tested experimentally and their Biological activities will be known (Ic50). With the help of Computational tools (QSAR), other compounds can be tested within a few minutes and their biological activities(Ic50) will be known without experiment . This has made QSAR time-saving,easy and cost effective.
(
The method of QSAR involves generating set of molecular properties of Compounds known as Descriptors. The Descriptors generated and the Ic50 of the Compounds tested experimentally will be used to generate a model in form of linear equation model (y= mx+c) where y is the biological activities of the unknown compounds. From the model generated, the Ic50 of more than 1000 compounds can be tested.
The Descriptors can be obtained from many Computer softwares. One of them is padel descriptors. Padel descriptors can generate more than 10,000 descriptors which can be used to predict new biological activities of Compounds.
The method is listed below
(1) Generation of descriptors: This can be achieved by using computer software such as padel descriptors software
(2) treatment of data: This involve elimination of unnecessary descriptors that may affect the result. This can be achieved by using data division software
(3) data division: this involve division of data into training and test set. The training set is used to build the model(equation) while the test set is used to check (test) the model built. This can be done by using data division software.
(4) Building of model: This is the final stage: The model can be built from software such as material studio.