You are here:
Harder, Faster, Better, Stronger: Towards a High-Throughput 3D Virtual Screening Triage Framework of the entire CompTox chemical inventory
Citation:
Goldsmith, M. Harder, Faster, Better, Stronger: Towards a High-Throughput 3D Virtual Screening Triage Framework of the entire CompTox chemical inventory. QSAR 2021 International Workshop on QSAR in Environmental and Health Sciences, Virtual, NC, June 07 - 10, 2021.
Impact/Purpose:
Abstract submitted to the QSAR 2021 International Workshop on QSAR in Environmental and Health Sciences June 2021. Here we demonstrate a tiered virtual workflow that enables the scoping of the toxicant-target paradigm more efficiently across larger chemical libraries and vaster target space. The demonstrated approach comprised of multiple generated target pharmacophore models, and 3-D conformational libraries of the CompTox dashboard inventory providing the largest/fastest and ultimately most efficient 3D multi-target virtual affinity fingerprint library scale-up to date.
Description:
The EPA’s computational toxicology dashboard chemical inventory currently contains nearly 880,000 chemicals. In vitro screening of this entire inventory would be a costly and lengthy endeavor, yet critical to further advance our understanding of molecular toxicology. The last attempted multiple-target virtual screen (DockScreen) was too computationally intensive (2 months on 21 processors for 1000 chemicals on 144 targets) making libraries on the size 1000 times larger challenging. Additionally, new vendor libraries of synthesizable chemistries (i.e. RealSPace or Enamine) or combinatorially screened ultra large DNA-encoded libraries (i.e. HitGen/Enko/XChem/WuXi) are in the billions of chemistries, making traditional virtual screening (docking) intractable and requiring tiered virtual screening workflows (pharmacophores). We developed 3D-DSSTOx, a conformational library of nearly 900,000 molecules and over 25 million 3D conformations for rapid pharmacophore screening, geometry-optimized and charged at its major charge state (neutral pH) in MOE 2020. Along with new automated structure-based pharmacophore generation tools we demonstrate the practicality of 3D screening ER alpha, AR, and TR across the entire 3D-DSSTox in under an hour. Here we demonstrate a tiered virtual workflow that enables the scoping of the toxicant-target paradigm more efficiently across larger chemical libraries and vaster target space. The demonstrated approach comprised of multiple generated target pharmacophore models, and 3-D conformational libraries of the CompTox dashboard inventory providing the largest/fastest and ultimately most efficient 3D multi-target virtual affinity fingerprint library scale-up to date. This abstract solely represents the views of the authors and not the view of the Agency.