Extended Data Fig. 2: Flow chart of single-particle cryo-EM data processing.

Micrographs (15,751 for antagonist-bound, 9,369 for ligand-free, 6,944 for agonist-bound Ste2) were collected from independent sessions on a Titan Krios operated at 300 kV. Each dataset was corrected for drift, beam-induced motion and radiation damage. After estimation of CTF parameters, particles were autopicked and subjected to two rounds of 3D classification. The number of good particles selected from each round of 3D classification is shown. The initial set of best particles was subjected to Bayesian Polishing, beam-tilt correction and per-particle CTF refinement. After detergent signal subtraction, 3D classification without alignment was performed. The resulting best class of particles was selected and subjected to Bayesian Polishing and per-particle CTF refinement. For ligand-free Ste2 and agonist-bound Ste2 datasets, non-uniform 3D refinement was performed after reverting detergent signal. For agonist-bound Ste2, a distinct inactive-like (grey) and active-like (sea green) state cryo-EM maps were obtained. Three additional classes that were heterogenous (pale green) and tending towards the active-like state were obtained (see methods). The global resolution of the final maps was calculated using gold-standard FSC of 0.143.