Most organic compounds are polymorphic, i.e. they can crystallize in more than one crystal structure. This explains why crystals of the same compound can exhibit different solubility and effectiveness as a drug.

Conventional polymorph prediction routines are based on thermodynamic stability and predict that only one form would exist, the `stable' form. However, due to nucleation kinetics, a thermodynamically unstable form may crystallize and prevent the stable form from crystallizing (Ostwald's rule of stages). Thermodynamics predicts subsequent conversion to the stable form, but as conversion in the crystalline phase is often very slow, unstable polymorphic forms do actually exist over extended periods of time. Therefore, unstable polymorphic forms cannot be ignored in industry and they can even be of use. To be able to predict their formation, we are developing a kinetic polymorph predictor that simulates the nucleation process using a Monte Carlo algorithm.[1]

To ensure computability, our kinetic polymorph predictor uses crystal graphs: abstract descriptions of molecular structure, which contain all the important aspects that determine nucleation. We have been using the molecular mechanics  program Cerius to compute crystal graphs from crystallographic data. In order to be independent of crystallographic data and to be more accurate, we started using the open source ab initio DFT program Siesta in combination with JuNoLo that adds a vdW-DF-functional. Together they function as a sophisticated energy-based polymorph prediction routine, generating many possible polymorphic structures. We added the functionality of representing them as crystal graphs.

Our kinetic polymorph predictor is highly automated and efficient, making decisions by root-finding methods rather than using grid methods. Thus, it is feasible to compare several hundreds of crystal graphs that result from the energy-based polymorph prediction routine.

Nucleation would be strongly influenced by the existence of precursor clusters in solution. We therefore added the possibility of simulating nucleation from all possible precursor clusters.

We will show results for some polymorphic drugs.

• Legal notice:
  

  Copyright (c) Pielaszek Research, all rights reserved.
  The above materials, including auxiliary resources, are subject to Publisher's copyright and the Author(s) intellectual rights. Without limiting Author(s) rights under respective Copyright Transfer Agreement, no part of the above documents may be reproduced without the express written permission of Pielaszek Research, the Publisher. Express permission from the Author(s) is required to use the above materials for academic purposes, such as lectures or scientific presentations.
  In every case, proper references including Author(s) name(s) and URL of this webpage: https://science24.com/paper/29621 must be provided.

1.  Complete deracemisation of proteinogenic amino acids using Viedma ripening
2. The role of liquid ordering in solution growth
3. High resolution protein crystals using an efficient convention-free growth geometry
4. Complete deracemisation of proteinogenic amino acids using Viedma ripening