Designing a candidate drug molecule

Two distinct things were done by machine, and it is worth separating them: picking TNIK out of the biology as a fibrosis target, and generating a small molecule that binds it. Everything after that was human and slow, the synthesis, the animal work, the regulatory filings, the recruitment across twenty two sites in China, and the reading of the results. The 60 mg arm gained 98.4 mL of forced vital capacity where placebo lost 20.3 mL, over twelve weeks in 71 patients, which is a signal worth a larger trial and not a demonstration that the drug works. What is contested is candidate generation. What is not is the demonstration that a molecule helps a sick person, which no model produces and which still costs years, money and volunteers.