The world of pharmaceuticals is changing rapidly, because of demographic pressures, social lifestyle changes,  growing economies in Asia and Africa, and technological developments in data storage, algorithms, and cloud computing. Culgi contributes in two different ways to the revolution. On the one hand, our models are used in late-stage pre-clinical research for material development (drug delivery systems). More recently, our models are also applied to early-stage drug discovery. In all cases, the calculations follow the same principle. We use coarse-grained simulations to calculate the thermodynamic fate of a potential drug in a complex environment, such a lipid membrane (transport), protein (docking) or tablet (formulation).    Of particular importance is the application of coarse-grained simulations to so-called biologicals where so far few if none method was available.

hERG paradigm

Drug discovery uses hERG interference as a means to predict whether a new drug could have harmful side effects. Until recently, there was no alternative but costly experimentation or uncertain QSAR. Since very recent (2017), a structural model for hERG protein is available, that allows now, for the first time, to develop a coarse-grained structure-based docking model. The naissance of the CULGI model is an exciting new development. The model demonstrates the maximum performance of all calculation modules in CULGI in joint action: ultrafast charge calculation, automation of fragmentation and parameterization, and coarse-grained thermodynamic integration. The coarse-grained binding model is the first of its kind. It fits right in between detailed but very time-expensive molecular models, and very rough high-speed QSAR methods: a new paradigm is born.


Polymorph Stability

A key issue in the formulation of drugs is the stability of the drug crystal in the environment of the formulation. In general, one wishes a fine dispersion of small crystals, that dissolve readily in the body. But this is not at all easy to achieve. For example, differential surface energies of crystal faces, lead to growth or dissolution into macroscopic needles, plates, cubes, or more complex shapes (so-called polymorphs).

The differential surface energies are dictated by the molecular structure of the particular cut through the crystal, and formulation excipient composition.

Another challenge is that crystals may agglomerate into less effective large colloid aggregates.

We have developed the first calculation method to study these phenomena with one unifying concept. The method is rapid taking 10 core-minutes or less per calculation. The method is also reliable since we calculate calibrated free energies, in combination with an experimental crystal structure.

The principle is highlighted in the picture. Given a crystal structure of the drug (cif file), we replicate the unit cell coarse-grained in all lattice directions. The depicted COSMO charge envelop is from CULGI’s ultrafast empirical charge calculation method. Fragments and parameters are assigned automatically using CULGI’s proprietary AFP algorithm. The coarse-grained nanocrystal is soaked in formulation solution, that can be of arbitrary composition, including alternative solvents (such as a few percents DMSO aqueous solution) polymers, surfactants charges and so on. Interfacial surface free energies of different faces are calculated directly, or the system is further used for a micro-scale hydrodynamics simulation.

The method replaces force-field based calculations, that are too slow and too unreliable anyway. Also, the application of QSAR to these systems is cumbersome because of lack of data, and because of the intrinsic three-dimensional nature of the crystal surface.

Yet the method is fast enough to be applied in an in silico screening over formulations.


TableT Stability

The formulation of a drug into a tablet is a non-trivial and expensive matter. One of the many aspects is the hygroscopy of polymer excipient and or drug, which could be a serious limitation to long-term storage. In this example a CULGI molecular model was used to drive the calculation of so-called Flory-Huggings parameters for water, polymer, and drug, In a second mesoscopic model, the parameters were used to calculate the ternary phase diagram. The depicted polymer and drug are simple and used for validation only.  In the actual application, much more complex polymers and drugs (kinase inhibitors) were being tested, using the same method.


Vesicular delivery

It is estimated that 1/3 of all modern drugs are nasty: they do not dissolve well and are very difficult to formulate in a homogeneous system. A neat workaround could be to deliver the drug through a so-called vesicular delivery system,  in which the drug is either incorporated in the bilayer or the central cavity. In this case study, we screened the interaction between several cancer medicins (kinase inhibitors), against various vesicle formulations. In all cases, we found that drugs accumulate in the bilayer.  In doing so, the drugs destabilizes the vesicle. As a consequence, the envisioned delivery system is in fact worthless. A negative result with significant impact.


Application of CULGI coarse-grained modeling to pharmaceutical problems is an area of active research. A tremendous potential is available in modeling biological aggregation, polymorph prediction, membrane permeability prediction, and of course structure based docking. Coarse-grained structure-based models put computational chemistry right in the seat of action, in between IP for drug discovery and IP for development.

Download a whitepaper

We have many more applications, in areas ranging from antibody aggregation, modified peptide solubility prediction, coarse-grained crystal structure prediction, and so forth. If you are interested in a white paper, please contact us.

Get in touch with one of our sales representives