As a leader in the field of hit identification, Creative Biolabs offers a full package of virtual screening services for preclinical drug discovery. Relying on our advanced technologies and rich experience, we can provide a powerful support vector machine (SVM) technique for potential drug candidate classification.
Support Vector Machines (SVM) is a powerful classification and regression tool in various machine learning applications. As a member of the unsupervised machine learning methods, it aims to create a decision boundary between two classes that enables the prediction of labels from one or more feature vectors. This decision boundary, known as the hyperplane, is orientated in such a way that it is as far as possible from the closest data points from each of the classes. A key feature of SVM is the attempt to minimize the error on training data and reduce the computational complexity of models to avoid overfitting by using the so-called structural risk minimization. It has been reported that SVM has a number of advantages over classical methods, such as stability, simple geometric interpretation, and use of kernels for nonlinear decisions.
Fig.1 Projection into high-dimensional feature space. Using a mapping function F, active (empty gray points) and inactive (filled pink points) compounds that are not Q7 linearly separable in low-dimensional input space (a) are projected into high-dimensional feature space (b) and separated by the maximum margin hyperplane. Points intercepted by the dotted line are called ‘support vectors’ (circled points). (Lavecchia, 2015)
In recent years, SVMs have become increasingly popular in multiple areas that are relevant to drug discovery. Among these, the prediction of novel active compounds represents the major application theme. For example, SVM classification can be used to distinguish active from inactive compounds via binary class label predictions. SVM can also be utilized to rank database compounds according to their probability to be active in virtual screening tasks. Furthermore, SVM can be applied to predict actual compound potency values as well as druggability scores for targets.
The traditional drug discovery process is time-consuming as it involves an iterative procedure of finding active compounds from a large collection of compounds. Working in the field of drug discovery for many years, Creative Biolabs has established an advanced technology platform to offer SVM technique which can aid the old screening process using the maximum margin hyperplanes. This hyperplane could separate the active from the inactive compounds and has the largest possible distance from any labeled compound.
As a leader in the field of virtual screening with great reputation and experience, Creative Biolabs is dedicated to taking the responsibility to promote your projects with the help of our advanced technologies and resources. For more detail information, please feel free to contact us. We will get back to you as soon as possible.
Reference
For Research Use Only.
