Our mission and vision

A lot of patients still suffer because of lack of effective drugs. According to our experience, most of diseases can be treated, at least at animal level, from even a relatively small library (~ 4, 000 natural products or active compounds). Given the vast chemical space, any disease, even for cancer, should find a cue or multiple cues. High throughput screening and target-based drug design have achieved big success. Yet there are a lot of room for improvement.

We combined the wet lab verification with recently fast developed deep learning to solving the problems in drug development. Deep learning (part of AI) provides us an overall and comprehensive understanding the similarities of chemicals, chemical-protein interactions, properties of chemicals, chemical gene relationships and etc. Thus, AI will eventually take over the responsibility for novelty disclosure, target finding, ligand design and etc. AI will bring more powerful pre-clinical candidates for further studies.

Company Profile

Our company aim at redefining the preclinical steps of drug development, from high throughput screening / CADD to data and AI driven strategies, that possessing the premise of high efficiency and intelligence. Specifically, we are working on: the drug-transcriptome and AI based efficacy prediction system (animal level testable); virtual screening system based on ligand-target binding data; prediction of chemical medicinal properties; target prediction system based on omics data, inversing from tool compound and experimental validation; design and optimization of lead molecules; prediction of toxicity and side effects. We aim to provide the highly efficient tools to facilitate big pharma for safe and effective clinical candidates.

Till now, our research paper upon deep learning based efficacy prediction system (DLEPS) has been accepted in Nature biotechnology. We also filed the patent on the architecture of small molecules – deep neural network – gene signatures – efficacy. Through DLEPS, we have found 3 effective anti-obesity molecules, one anti-hyperuricemia molecule with anti-inflammation and fibrosis benefits, 4 anti-NASH molecules with novel targets and MOA. The overall accuracy of DLEPS at animal level is around 60%.

AI platform

AI covers target finding, ligand design, retrosynthesis and etc., of our pre-clinical research.

learn more

Pipelines

We arranged pipelines in aging-related diseases, metabolic disorders and cancers.

learn more

Big Data

Big Data fulfill the AI's requirements need to be generated for exploring AI's potential.

learn more

Wet Lab

We have built chemistry lab and animal facilities to support the innovative findings.

learn more

Latest News

Investors