Features

CULGI software is one of the very few packages available in the world that encompasses an integrated suite of all relevant computational chemistry algorithms, from quantum to molecular and coarse-grained modeling, to informatics and thermodynamics, and all sorts of automated mappings between different levels.

A unique and compelling feature of CULGI software is the concept of scripted workflows. Workflows can be edited through either our proprietary graphical scripting editor or python.

In a typical application, a bespoke script is made by either client or us in a service contract.

Scripting is the key to success, in modeling a wide variety of systems.

Quantum Chemistry

CULGI includes a variety of empirical, semi-empirical and ab-initio calculation algorithms. Some of the algorithms are from NWChem and accessible through an interface. Other algorithms are implemented by us (AM1, and Qeq).

We have optimized Qeq for database screening. For a given molecule of MW 500, a CULGI charge Qeq calculation would take a few seconds, a semi-empirical calculation a few minutes, and a DFT calculation a few hours or days.  The parameter sets of all CULGI methods are open and can be modified as suited by the user. For example, a proprietary optimization of Qeq calculations as input to COSMO-RS calculations is entirely possible, by writing a script that loops over the CULGI database.

Through the interface to NWChem, CULGI can model the quantum behavior of solids (plane wave)  and molecules. The range of Density Functional methods in NWChem is comprehensive and covers all popular functionals.

NWChem is included in the installation.

Molecular Dynamics

CULGI has an extensive suite of molecular dynamics algorithms, including a 3D molecular sketcher, a powerful atomtyper, and amorphous builder. CULGI also has the capability to find new torsion force field parameters automatically. CULGI can read and write all popular molecular modeling formats through Open Babel. Even entire simulations can be imported and exported (on script level), for example to LAMMPS. The multiscale approach is also instrumental in building original structures. For example, in a typical scenario, one would first establish a polymer system coarse-grained. After relaxation and subsequent reverse mapping (coarse-grained to molecular), one would then continue in molecular dynamics. In this way, very complex initial structures can be created.

Unity in Diversity

CULGI software is a general modeling package, and, by means of our scripting editor, can be used just as easily for providing solutions in Oil and Gas, Personal and Home Care, Chemicals, and Pharma.

 

Coarse-Grained Simulations

CULGI is the world leader in providing modeling solution for nano- and microstructured chemical systems, whether from a synthetic or biological origin. The principal method is a set of proprietary algorithms that enable the automatic coarse-graining of any chemical into fragments. The method is referred to as Automated Fragmentation and Parameterisation.

Coarse-grained simulation engines include Brownian and dissipative particle dynamics. Interaction models are comprehensive. They can consist of actives, charged systems, colloids, polymers, surfactants, and also reactive systems. Coarse-grained forcefields can be simple linear as in dissipative particle dynamics, or more complex ones (by using tabular input).

Chemical Informatics and Machine Learning

CULGI has an extensive database of pre-calculated molecules. The database serves as input for industrial projects. We also use the database to calibrate our chemical informatics algorithms. CULGI has a comprehensive set of all useful descriptors known to date. Third party software can further extend the set of descriptors, through CULGI’s open data structure and python interface.

CULGI includes a range of statistical calculation methods, such as linear and non-linear regression, and genetic algorithms.

One of many ways one could extend the modeling capabilities is the interfacing of CULGI to Machine Learning software, for instance, TensorFlow (from Google).  By importing both CULGI and TensorFlow in a python script, one enables the combination of state of the art physics-modeling with state of the art data-driven modeling.

Engineering ThermodynamicS: COSMO-RS

CULGI includes a re-coded and re-optimized version of the powerful COSMO-RS methods (invented by Dr. Klamt, COSMOlogic Gmbh) .

Using COMO-RS, the user is capable of predicting thermodynamic properties such as pK, logP and mixing Gibbs energies from the molecular composition.

COSMO-RS is particularly useful for calibration coarse-grained simulations.
Another special feature of CULGI’s COSMO-RS is that we can offer the method based on superfast empirical and semi-empirical quantum calculations.

Mesoscopic Modeling

CULGI has cutting-edge methods for mesoscopic dynamics simulations of complex polymer systems, which are unsurpassed in the world. These so-called dynamical density functional methods (or generalized phase-field models) allow the efficient calculation of sub-micron structures in composites, blends, and formulations.

Engineering ThermodynamicS: UNIFAC

UNIFAC is the de-facto standard in chemical engineering thermodynamics. It is used to calculate a variety of thermodynamic properties of industrial solvent (mixtures), such as mixing Gibbs energies. CULGI has re-implemented the UNIFAC method from NIST, based on the extensive Thermodynamic Data Engine. The implementation uses CULGI’s proprietary SMARTS-based group-typer that allows the rapid identification of user-supplied groups or fragments in a molecule. UNIFAC calculations are high-speed. In a typical situation, the calculation of an activity coefficient takes less than a millisecond per system.

MAPPING

The robust mapping between different levels of modeling (quantum, molecular, coarse-grained, mesoscopic, macroscopic) is instrumental in an industrial environment. Since the founding of CULGI, we have been inventing methods for all kinds of mapping: from quantum to molecule, from molecule to coarse-grained and back, from coarse-grained to mesoscopic, and so forth. In a typical project, we first use Automated Fragmentation and Parameterization to identify groups and parameters based on a quantum and molecular model (mapping quantum to molecular, to coarse-grained). Then the coarse-grained models are used to build a 3D system. The coarse-grained system could then be mapped back to a molecular force field, or further coarse-grained to mesoscopic (coarse-grained to mesoscopic), or used as is.

Get in touch with one of our sales representives