A worldwide team of scientists and programmers joined forces in a blind test to spot the best computer tricks for hunting coronavirus medicines. They used secret data from a big AI project that looks for drugs against SARS‑CoV‑2 and MERS‑CoV enzymes. The challenge asked participants to guess how strong each small molecule would be and predict its exact shape inside the target protein. They also had to estimate how the drugs behave in the body, like absorption and safety.

The competition was run openly. No one knew the correct answers until the end, so every team had to rely on their own models. After all entries were scored, a leaderboard showed who did best at each task. The organizers then dug into the results to see what methods worked, where mistakes were common, and how future challenges could be better designed.

One key lesson was that the newest platforms, such as Polaris, make it easier to set up fair tests and let many people try their ideas. The data and the scoring system were shared with everyone, so anyone could replicate or improve on the work. This openness helps build trust in machine‑learning tools for drug discovery.

The challenge also produced a set of papers that dive deeper into the winning approaches. Together, these works aim to make computational drug design more reliable and encourage the scientific community to keep testing new ideas in blind, unbiased ways.

Global Hack to Find Coronavirus Drugs

Actions