High throughput quantum chemistry for molecular design
ConstruQt is our core molecular design tool that underpins all other automation and machine learning technologies that ChemAlive is developing. It is currently available as an API and allows for quick deployment of quantum chemical calculations to enhance your chemistry in the following ways:
- Transforms list of SMILES molecular designations into state-of-the-art 3D molecular structures in SD format
- Manages the conformational space of the molecules with a robust shape searching algorithm
- Generates all reasonable tautomeric forms of the molecule and prioritizes them by energy
- Generates all stereoisomeric forms of the molecules and differentiates them by energy
- All molecules are stored in our unique database (Qontext) architecture making the calculations easily augmented and carried through to other processes
- We currently maintain the largest quantum chemical database known. We are building a contextual data structure commensurate with advanced machine learning algorithms. It is called Qontext.
Try ConstruQt Interface
Most conformations are energetically accessible
Distribution of conformational relative energies (kcal/mol) of over 1.5 million molecules and 35 million conformations at the PM6 semi-empirical level contained in our Qontext database
Classical Force Fields Are Unreliable
The energetic error between universal force field (UFF) relative energies and PM6 derived free-energies (kcal/mol) as a function of PM6 conformational relative energy. Notice the UFF error is upwards of 80% and increases as we move away from equilibrium
Why high throughput quantum chemistry?
- ConstruQt provides a sea-change in accuracy and scope for molecular modeling
- Moving away from alchemical rules-based prediction, ConstruQt computes molecular energies and electronic properties from fundamental physics
- With quantum chemistry you get not just what is possible, but what is probable
- ConstruQt allows the deployment of these vastly improved metrics at the scale that you are used to using with cheminformatics and classical mechanics force field calculations
- The core ConstruQt computational approach centers around cheminformatics management, classical mechanics initial guess and follow-on semi-empirical quantum chemical calculations for energy assessment
- Density functional theory calculations can be launched as well.
- ConstruQt is available as an API to connect to the ChemAlive AWS Supercomputer
- With Dynamic scalability, ConstruQt can handle up to 500 unique molecules per minute per server (depending on size and rotational degrees of freedom)
- The engine accepts SMILES as input and returns computed 3D data in SD format
- Whether it be for your chemistry SaaS, on-line catalog or virtual library, the API delivers far more accurate data at unmatched speeds and accuracy.
- Conformations - Molecular shape is 70% of drug discovery and key to all follow-on calculations at more accurate levels. Legacy methods are unable to accurately predict molecular shape
- Tautomers - Fundamental to acidity constants and binding affinities, you need to know which tautomer is most stable to understand your structure based determination
- Diastereomers - Diastereomers can differ sharply in energy. When a set of stereoisomers can result from a reaction, ConstruQt can help you begin to identify which one is present. Follow-on calculations of optical rotation can further elucidate the chemical structure.
- Micro-states - We can accurately predict protomers using automatic linear regression analysis (coming soon).
- The ChemAlive postgreSQL database, Qontext, is built by chemists for chemists
- The API additionally allows access to the largest repository of quantum chemical data maintained in the ChemAlive database
- ConstruQt allows direct access to this high quality data by library or single molecule searching
- Submissions to Qontext will contextualize molecular structure based on the computations requested.