Molecular mechanics force fields of drug molecules are an essential component for many structure-based drug design predictions. Force fields provide an accurate description of intramolecular and intermolecular interactions by parameterizing a functional form to characterize the potential energy of molecules. At present, there are still some common issues in small molecular force fields, such as unsatisfactory accuracy and precision, insufficient chemical space coverage, and limited capability for customization. In addition, the general molecular mechanics force field has a tedious development process that requires many cumbersome human interventions that prevent its fast extension to new sets of molecules.
We have developed and improved XForce Field to address these common issues. Our proprietary next- generation general molecular force field for drug discovery and development. Set below are notable features of our XForce Field.
XForce Field has a comprehensive chemical space coverage stemming from a variety of different training sets. The number of training compounds included in the development of XForce Field is around two orders more than the public academic force fields. It facilitates the exploration of chemical space and improves the success rate of drug design.
XForce Field is trained using quantum chemistry and experimental data, and thus provides a reliable presentation of molecular conformations, single molecule properties, intermolecular interactions, and molecular behavior in solution and in drug targets. Compared to the results of high-accuracy quantum mechanics calculations, the accuracy of our customized force fields can reach the level of 3-6 kJ/mol, representing a much higher accuracy compared to that of commonly-used general force fields.
XForce Field supports both cloud and local deployment to cater to users' specific needs. The cloud deployment provides users with access to computing resources that enable a fast verification and re-parameterization process. On the other hand, the local deployment may be more suitable for developing customized molecular force fields based on users' particular internal data.