AI based Lead Compound Detection
We have proprietary AI platform to train the algorithm on given compound library and test it for new molecule’s activity. This platform uses deep neural network and decision tree technique together to predict the activity of compound (IC50, Ki). We have high performance computing access that allow us to use these algorithms extensively and detect the best compound. We use comprehensive list of 1D, 2D and 3D chemical features of molecule in training process. In addition, we also use convolution neural network for the 2D image of compound in order to improve the accuracy of result.
Virtual Compound Screening
We have 13 million virtual compound library that has been curated from different sources (literature and databases). We perform virtual screening of these compounds against given protein target and arrange these molecules based on their individual binding affinity. We also provide complete report for 10 best molecules including their reported toxicity and patent information. Our scoring function is geometry and chemical properties complementary based method that screens the library in rapid manner.
We provide extensive services on docking of small molecule with the given protein target. We used our proprietary and other best available public and licensed docking algorithm. We present a detailed comparative report of docking using various applications and report the best compound. In this service we include MMPBSA binding energy of best selected pose to further verify the complex. This report also includes the detailed information on type of bonds and residues involved in the protein-ligand interaction.
Molecular Dynamic Simulation
Dynamic behaviour of protein/protein-ligand system can be explained by running molecular dynamic simulation. We provide complete solution in MD simulation for protein conformational behaviour, receptor-drug interaction, and other events that require the systematic evaluation of molecular properties in dynamic molecular systems.