Protac Puzzle

This journal cover was inspired by the work presented in the article authored by Nils Dunlop, Francisco Erazo, Farzaneh Jalalypour, and Rocío Mercado at Chalmers University of Technology. Their study looks at how computer models can predict the way three molecules come together to form a complex structure called a PROTAC ternary complex. PROTACs are special types of molecules that can tag unwanted or harmful proteins in our cells for destruction. This helps scientists find new ways to treat diseases by removing problem proteins instead of just blocking them.

In this work, the team used two powerful artificial intelligence systems—AlphaFold3 and Boltz‑1—to predict how these large, flexible PROTAC molecules connect two proteins at once. They tested both models on 62 known complexes from scientific databases and found that AlphaFold3 did a better job placing the PROTAC in the correct position. By including the actual shape of the PROTAC molecule in their calculations, they could more precisely model the way the three components interact.

To make it easier for others to explore this process, they also built a platform called protacfold.xyz that automates how these complex predictions are generated and compared. The study shows how digital tools can help understand molecular teamwork inside cells and speed up the design of new protein‑targeting medicines.