Extended Data Fig. 9: Workflow of cryo-EM data processing of PI3Kγ·ATP derived from an un-tilted dataset of PI3Kγ supplemented with Gβγ and ATP.
From: Molecular basis for Gβγ-mediated activation of phosphoinositide 3-kinase γ
We collected 3936 frames for the PI3Kγ sample supplemented with Gβγ and ATP (white scale bar = 100 nm in the micrograph presented in the top left panel) and processed the data in CryoSPARC, resulting in 472,023 particles after several rounds of 2D classification. The 2D classes showing distinct Gβγ density are indicated by red boxes. The 2D classes showing no Gβγ density are indicated by yellow boxes. The classes in green boxes are undetermined. We generated 2 ab initio maps using the particles from the “red” and “green” 2D classes, followed by heterogeneous refinement in CryoSPARC, resulting in a “junk particle” and “Gβγ-bound PI3Kγ” cryo-EM maps. In parallel, we processed the particles from the “yellow” and “green” 2D classes using the same strategy, resulting in a “PI3Kγ alone” and another “junk particle” cryo-EM map. We then used the 3 maps (1: junk-particle, 2: PI3Kγ–Gβγ, and 3: PI3Kγ alone) for sorting all particles using heterogeneous refinement, and then did homogeneous and local refinements for the “PI3Kγ alone” data, resulting in a 3.3 Å cryo-EM reconstruction of PI3Kγ·ATP (Extended Data Fig. 2E).
