Fig. 2: Distributions of molecules generated by Lingo3DMol, Pocket2Mol and TargetDiff on DUD-E targets (nā=ā101).
From: Generation of 3D molecules in pockets via a language model
The drug-like region with QEDāā„ā0.3 and SASāā¤ā5 is indicated with red boxes. aāe, Heatmaps show the distribution of key properties for generated molecules, including count (a), molecular weight (b), minimum in-place GlideSP score (c), GlideSP redocking score (d) and r.m.s.d. versus low-energy conformer (e). These distributions are depicted along SAS and QED.
