Publication details
Prediction of QM-like partial atomic charges for alphafold
| Authors | |
|---|---|
| Year of publication | 2022 |
| Type | Appeared in Conference without Proceedings |
| MU Faculty or unit | |
| Citation | |
| Description | The basic functional unit of all living organisms are proteins. If we want to understand the function of a protein, it is essential to know its structure. Thanks to the AlphaFold algorithm [1], which predicts the structure of proteins based on their sequence, the number of predicted structures is growing very fast. However, due to the high computational complexity of the quantum chemistry apparatus, we are not able to calculate for such large structures another key characteristic, i.e., electron density. A suitable approximation is the concept of partial atomic charges, which describe how much electron density belongs to each protein atom. The most accurate but in our case computationally inapplicable way to obtain partial atomic charges is to derive them directly from the electron density. A significantly faster alternative is to use one of the empirical methods that use only the coordinates of the atoms and information about the bonds between them for the calculation. However, empirical methods must go through a parameterization process, where parameters are searched for such that the resulting empirical partial charges are as similar as possible to those from quantum chemistry. This paper presents a new empirical method called Split-charge equilibration with parameterized initial charges (SQE+qp) [2] adapted for the AlphaFold Protein Structure Database. Our method can reproduce the partial atomic charges of QM with high accuracy. We also present an implementation of SQE+qp and its parameters via the Atomic Charge Calculator II [3] web application at https://acc2.ncbr.muni.cz. Thus, we provide the scientific community with a freely available online tool for calculating QM-like partial atomic charges. |