This type of search is a combination of 2D Fingerprint, 3D Shape, 3D Electrostatic, and 3D Shape Electrostatic. It searches by all of them and ranks the results based on equally weighted similarity scores.
2. 2D Fingerprint (Read More)
Description: AI-Learned Tanimoto similarity of Morgan 2048 bit fingerprints with radius 2.
Fingerprints are created by applying a kernel to a molecule and are bit vectors. The presence or absence of a specific fragment is encoded in each bit.
3. 3D Shape (Read More)
Description: AI-Learned Shapesim similarity of aligned conformers volume overlap (for training, one pair of molecules overlap of 10x10 conformers was computed and best score was used as the value for training).
Shapesim computes the Shape Tanimoto Distance, it is useful when you want to obtain the similarity of 3D molecular shapes instead of partial charges based on graph structure of molecules. This method calculates the overlap integrals of the Vander Waals surface of two molecules.1 Figure is taken from Kumar, Zhang 20182.
4. 3D Electrostatic (Read More)
Description: AI-Learned Espsim similarity of aligned conformers electrostatic surface overlap (for training, one pair of molecules overlap of 10x10 conformers was computed and for the best shapesim aligned conformers electrostatic surfaces were integrated and this value was used for training).
Espsim calculates the overlap integrals of the electrostatic potentials (generated from Coulomb potentials) of two molecules or fragments.1
5. 3D Shape-Electrostatic (Read More)
Combined 3D Shape and 3D Electrostatic similarity.1
On the Value of Using 3D Shape and Electrostatic Similarities in Deep Generative Methods Giovanni Bolcato, Esther Heid, and Jonas Boström Journal of Chemical Information and Modeling 2022 62 (6), 1388-1398 DOI: 10.1021/acs.jcim.1c01535 ↩↩↩
Kumar, A. & Zhang, K. Y. J. Advances in the Development of Shape Similarity Methods and Their Application in Drug Discovery. Front Chem 6, 315 (2018). ↩