From target identification to finding candidates of a target, coincided with technological surge of computational analysis methods and power, Artificial Intelligence (AI) started to take part in processes that people have reviewed, investigated, and performed in the course of developing new drugs. Just as computer vision enhanced productivity and quality of production industry, the AI-driven revolution along the prelude of new drug development is awaiting. DeepHits™, an AI-based drug development platform brand of JLKBIO, was devised and developed to augment conventional drug development processes.
Our first platform DeepHits Generation focuses on de novo candidate design and molecular optimization. Although many AI companies have platforms that have the same purpose as us, our differentiation stands out from two important technical advantages. Firstly, we reached forward to maximizing diversity of compounds by using various data as well as deep learning models that apply different phenotypes of compounds (SMILES, Graph, Inchi, etc.). Secondly, we established efficient and reliable way of generating chemical of interest. Unlike other approaches, DeepHits Generation uses Drug Discovery Space to independently predict the desired characteristics and create compounds that are similar to the values of the characteristics set based on the predicted values. Most importantly, because the generation process is independent of the characteristic prediction process, it can be applied to the generation process by selectively specifying the desired characteristics depending on the situation.
Just as there are experts in each field of drug development, we designed differently according to its purpose. Starting from the introduction of DeepHits Generation, in near future, we are looking forward to release DeepHits Toxicity, DeepHits ADME, and DeepHits Repurpose for toxicity prediction, drug profiling, and drug repurposing, respectively. Thus, according to the needs of our partners or customers, JLKBIO is able to provide appropriate answers.