Entos to Develop Open-Source Machine-Learning Tool for Chemical Discovery

New Partnership to Create Innovative Platform for Research and Education

November 30, 2020

Many fields of research, from biology to medicine to materials engineering, are, at their most basic level, a study of how atoms and molecules interact with each other. Those interactions determine how a protein operates in a cell, whether a potential drug molecule will be effective, how a new polymer will behave, and much more.

Traditionally, research in those areas was done using trial-and-error experimentation. Scientists would create new materials in a lab and then test them to see if they had desirable properties. However, with the advent of quantum chemistry, which examines molecules through the lens of quantum mechanics, researchers can accurately predict the properties of a new molecule or material before it is even made.

Entos Inc., a Pasadena-based startup, aims to provide tools for molecular and materials discovery, through the use of physics-based machine learning. Now, Entos has inked a new relationship that the company's founders say will help it expand the reach and impact of its current technology. That collaboration, with Schmidt Futures, a philanthropic initiative co-founded by Eric and Wendy Schmidt, provides Entos with nearly $2.5 million in funding from The Eric and Wendy Schmidt Fund for Strategic Innovation to develop a platform for chemical discovery based on machine-learning technology to advance both chemical research and education.

The goal for the new platform, called Entos Envision, is to combine machine-learning with quantum mechanics to enable researchers to find desirable molecules as easily as Google Maps makes route planning for motorists.

In July, Entos unveiled OrbNet, a machine-learning technology that performs quantum-chemistry predictions over 1,000 times faster than was previously possible. That leap in speed means that such predictions can be done in real time. Before the debut of OrbNet, those predictions could take as long as several days, greatly slowing the progress of research.

Fundamentally changing the speed of accurate molecular predictions provides the opportunity to revolutionize the way that chemistry software works, both in discovery and education.

Entos Envision will be an open-source platform that brings rigorous quantum mechanics and machine learning together to enable chemists to predict molecular properties and to discover promising new molecular systems, without the need for expert computational chemistry training.

It is part of our core values at Entos to provide products free for academic research and for education, and we’re excited to partner with Schmidt Futures to help that to happen. Our mission with Envision is simple: We want everyone who is thinking or talking about molecules to have Envision in their hands.

The open-source ethos and goal of providing new tools to researchers in higher education, meshes well with the mission of Schmidt Futures, says Stu Feldman, chief scientist of Schmidt Futures.

We are excited to support this project that combines the latest research in computational chemistry and a plan for an open platform that can radically improve chemistry education and support researchers who need novel compounds. This style of innovation and broad impact is precisely what we seek to fund.

Miller says the grant from Schmidt Futures will allow Entos to move forward rapidly in developing Envision, with the opportunity to broadly and positively impact the chemistry community. Already, Entos Envision is appearing in classrooms at the California Institute of Technology and the University of Bristol, with plans to rapidly expand to other schools and universities in the coming months.

“Partnering with Schmidt Futures, we have the ambition to create a tool that will benefit every chemistry student and researcher, regardless of location or access to funding,” Miller says.