ITMO Scientists Create Fast and Accurate Tool for Calculating Cavity Volumes in Calixarenes

To be delivered to the desired area of the body, an active substance needs to be encapsulated into container molecules. The same principle is used to produce  supramolecular catalysts and sorbents for gas capture and absorption. For a guest molecule’s  encapsulation to be successful, scientists need to know the precise shape and volume of its intramolecular cavities. This is typically done using specialized software that simulates the rolling of virtual probe spheres along the outer and inner surfaces of host molecules. The method, however, performs poorly with calixarenes or macrocycles, which are vase- or cup-shaped and have unclear boundaries due to protruding external atoms. Current calculation methods for such molecules require time-demanding hand-tuning. 

A group of ITMO researchers solved this problem by proposing two novel geometric approaches based on computational modeling and divide-and-conquer algorithms that can construct a closed shell for processing of calixarene molecules. The method allows specialists to calculate cavity volumes faster and more precisely than with existing solutions. 

The scientists applied two approaches when building closed shells. In the first case, they manually removed excess atoms outside the cavity and created a 3D model of a molecule; in particular, they loaded the coordinates of its atoms into AutoCAD to map their chemical bonds. Based on this data, they formed a convex mosaic shell of triangles, transformed each triangle into a plane surface and each structure – into a solid body; then, they subtracted spheres that correspond to the molecule’s atoms and calculated the volume of the resulting solid. For the second approach, employed in CaviDAC, the researchers needed to remove excess atoms and load molecule coordinates into the software; then, using the QuickHull algorithm, they built a closed shell that covered the molecule and calculated cavity volumes.

“Our approaches show good results and make calculations faster: it’s 5-10 seconds for CaviDAC compared to 1.5-2 hours (depending on a molecule’s geometry) if using 3D modeling,” highlights Sergei Karalash, the first author of the paper and an engineer at ITMO’s Infochemistry Scientific Center. 

The developed tools were tested on calixarenes and crown ethers – macrocyclic compounds that capture and retain cations of various metals. For that, the scientists calculated the volume of their cavities using the developed software and then compared their results with theoretical calculations and experimental data. The ITMO-developed software showed an average deviation of about 11%, while its analogs POVME, pywindow, MoloVol, and pyKVFinder reached up to 28-59% on the same front. Both approaches demonstrated a similar performance – their results differed by less than 3%. 

“If we know the cavity volumes of a host molecule and the full volume of a guest molecule, we can theoretically predict which guest molecules will better link with calixarenes and then experiment with predicted molecules. It’s much faster than via manual methods. Having accurate calculations of cavity volumes, we can predict the efficacy of guest molecule encapsulation when creating novel sorbents, supramolecular catalysts, and target drug delivery systems,” notes Anton Muravev, the lead author of the paper and an associate professor at ITMO’s Infochemistry Scientific Center.

Next, the authors plan to automate the deletion of excess atoms from 3D models of molecules and introduce new software features – for instance, recommendations on molecules to be encapsulated into calixarene cavities. 

The software’s code and files used for testing CaviDAC can be found here.