Episode 19: I Can See You
Travels, US election results, preprints & SPOILER ALERT: You Can See Everything discussion at the end of the pod.
Preprints:
Other Links:

The CYP challenge drew over 6,000 submissions. CYP2D6 was hardest to predict, and structure entries usually found the binding site but placed the ligand wrong.
By Maria Castellanos, Jon Swain and Hugo MacDermott-Opeskin. Read the full post on the OpenADMET blog.

Many CYP3A4 entries have ligands poorly supported by density, and Nikhil lists which to exclude when training models.
By Nikhil Gupta. Read the full post on the OpenADMET blog.

Taylor shows that GOODVIBES recovers the correlated motions of whole proteins in a simulated lysozyme crystal, matching what the MD trajectory says directly.
By Taylor Womack. Read the full post on the Prism blog.

Pooling co-folding poses and rescoring them with physics helped in the PXR structure track, but out-of-distribution ligands stayed hard.
By Maria Castellanos, Jon Swain and Hugo MacDermott-Opeskin. Read the full post on the OpenADMET blog.