About Us
Training the next generation of AI drug design with living data
Phase I clinical trial success rates have improved from ~50% to ~85% owing to advances in computational drug development tools.
Phase II clinical trial success rates have not improved substantially.
Most AI drug design platforms train with fundamentally limited and
biased data. Data that does not measure drug efficacy in a living system. Most have no biological validation layer whatsoever.
Biological efficacy is the only data type able to bring a drug to clinical use.
Azor Biotek's novel AI-guided drug design platform uses purpose built high throughput robotics to capture training data from live zebrafish and human xenograft models. We are building the viable alternative to rodent testing regulators have been searching for — vertebrates with human organs.
Zebrafish
Zebrafish have been involved in bringing dozens of drugs to clinical trial (e.g. ProHema & Fenfluramine)
Zebrafish screening naturally selects for therapeutics with high bioactivity, and allows for in vivo quantification of concrete physiological hallmarks.
See our background video on what makes zebrafish so amenable to high throughput drug screening. Compared to mice zebrafish evaluation is <1% the cost, <1% the time, uses <0.001% the material, is less invasive and more informative. With zebrafish it is possible to image living data in real time
