-
As an alternative to exhaustive searching, our fully automatic 4Dshape+ Similarity & Docking modules use either multiple binding poses generated by MD simulations of a scaffold and/or superposition of multiple scaffolds extracted from X-ray structures.
-
The docking algorithm attempts to resolve ligand-receptor steric clashes by identifying the optimal initial binding pose of the ligand which is further relaxed using the hybrid QM-MD simulations.
AI-Drug Discovery Platform
Our unique computational platform integrates Machine Learning techniques, 4Dshape+ similarity/Docking modules and hybrid QM/MD simulations to accelerate drug discovery.
Our computational solution AI-4Dshape+ includes the state of the art techniques of machine learning algorithms, Multiple ligands shape-based Similarity & Docking, and hybrid QM-MD simulations.
The AI-4Dshape+ software package features:
​
-
Fully automatic workflows.
-
Improved scaling, supporting thousands of compute resources (CPUs/GPUs).
-
4Dshape+ Similarity & Docking modules support multi-threaded parallelism, speeding up the search.
-
Provides new chemical starting points for novel targets or design optimized molecules.
-
Previous features described in 4Dshape+ Sim & Docking modules.
-
AI platform using multiple optimized and prebuilt machine learning workflows (KNIME).
-
Plugin Interface to Datawarrior (openmolecules.org/datawarrior/)
-
Virtual screening workflow using AMBER22 & AmberTools22 software packages. (ambermd.org/AmberTools.php)​
4Dshape+ Sim module
​
​
​
​
​
​
​
​
4Dshape+ Docking module includes MD-QM/MM-GBSA simulations/Binding Free energy calculation
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​​
​​
AI machine learning module, includes 6 machine learning algorithms.
​
​
​
​
​
​
​
​
​
​
​
​
​
​​​
- The search freedom using Multiple queries Docking process is greatly reduced compared to that in full-flexible docking and exhaustive search methods.
-
The results are ranked according to the docking score or QM/MM-GBSA Binding Energy to identify highly active compounds.
-
The multiple ligands shape-based screening of large ligand libraries can identify more effective active compounds with favorable pharmacokinetics and low cytotoxicity.
