Molecular Quantum Solutions (MQS) was founded in December 2019 by Mark Nicholas Jones (CEO) and Lukasz Ruszczynski (CTO) who know each other from their EU funded PhD projects (https://cordis.europa.eu/project/id/675251) at the Technical University of Denmark (DTU). During the final project conference they had the idea to start MQS to use the acquired knowledge from their PhDs, combine it and implement a full fledged software as a service solution for the pharma, biotech and chemical industries.
Grown tired of using inflexible software in the chemical engineering domain which feel outdated in comparison to many tools in signal processing and artificial intelligence applications, Mark and Lukasz want to deliver modern and future computational tools to accelerate research & development efforts by the pharma, biotech and chemical industry. Their tools include the use of high-performance computing and quantum-computers with quantum chemistry models and machine learning to calculate the properties of materials and chemicals in a fast and efficient way.
Users are able to screen for example new materials for batteries, more environmentally friendly solvents, new drugs and bio-degradable plastics before performing costly experiments in the laboratory.
“An important aspect is to contribute back to open-sourced and freely licensed scientific software for us”, says Mark. “As seen with many other start-ups and companies in IT, wrapping tools, automating the data pipeline and patching the algorithms with the newest research discoveries is the way to go if one wants to sustain and be flexible to react to new developments and trends.”
Mark and Lukasz standing on the rooftop in Copenhagen. If the cloudy weather in Copenhagen has given them the inspiration to work on a cloud based SaaS business might be a question worth to ask.
Scientists and engineers often fail to develop and simulate new processes and products because:
The experimental data for specific molecules are not available and have to first be measured and can become infeasible for a large number of molecules since this would be too time- and material-consuming.
No property model has been provided within the tool set of the users workplace IT infrastructure and if so, might lack the needed accuracy.
Further, an important aspect is to be able to perform a risk-assessment with the uncertainty of the property predictions and the final results of process or product calculations. Many prediction models in current tools do not provide any uncertainty bounds.
MQS has solutions to all of these problems. Their software tool provides quantum chemistry-based prediction models for thermodynamic properties with uncertainty provision and the capability for the user to include an already existing experimental data set which will tune the model for a specific chemical class. Due to the capabilities of today’s computers and server providers, quantum chemistry models can be run on multi-core systems for small and large molecules. And here comes the expertise of MQS: the quantum chemistry models have been carefully selected with respect to the application domain, to have the necessary trade-off between accuracy and computational resources.
“It doesn’t make sense to use the most accurate quantum chemistry model if you want to screen thousands of molecules, this would blow up the computational effort and cost way too much electricity. The importance is to select the correct model dedicated specifically for the chemical system you want to model and take the upper layer application into account which uses these property estimates to do for example a ranking of the chemical candidates” says Lukasz.
“Another example would be distillation, extraction or chromatography columns, in general separation processes, where for example phase equilibrium data of the individual compounds is needed," Mark adds. "With quantum chemistry one can fine-tune existing models which have been generated by a company and would like to evaluate a column design for a new chemical system for which they don’t have any data or they want to evaluate new extraction solvents which possibly perform better and are more sustainable.”
In general, property prediction is needed for almost everything, to simulate unit operations and flowsheets, formulation of products, finding new catalysts or bio-degradable plastics. The list goes on and on.
INAM met MQS during the #Hackcorona hackathon where they joined hundreds of other like-minded individuals, innovators and creative thinkers to find solutions to problems caused by the recent pandemic. A few months later, MQS was selected as a finalist for the first ever AdMaCamp program - the Advanced Materials Online Bootcamp - organised by INAM.
In more recent news, MQS has just scaled their team and are working hard to release the first demo version of their software this fall and are collaborating with several mid-sized and stock-exchange listed pharma and chemical companies.
For more info, visit their website http://mqs.dk/. If you'd like to get in touch with MQS, reach out to us at info@inam.berlin and we'd be happy to put you in contact.