AI Innovation Learning Machine for Development of Chemical Library for Drug Research

Purdue University scientists use a machine learning model to create new options for drug discovery pipelines. Credit: Purdue University / Gaurav Chopra

One-step multicomponent reactions with interpreted machine learning innovations for the development of chemical libraries for drug discovery.

Machine learning has been widely used in chemistry for the design of drugs and other processes.

Models that are prospectively tested for new reaction results and used to improve human understanding to interpret the chemical reactivity decisions made by such models are very limited.

Purdue University innovators have introduced chemical reactivity charts to help chemists translate reaction results using statistically powerful machine learning models that are trained in a number of small reactions. This work was published on Organic Letters,

“Developing new and rapid reactions is crucial to the design of chemical libraries in drug discovery,” said Gaurav Chopra, assistant professor of analytical and physical chemistry at Purdue’s College of Science. “We have developed a new, fast multicomponent reaction with one pot (MCR) of N-sulfonylimines are used as representative cases to generate training data for machine learning models, predict reaction results and test prospects blindly prospectively.

“We hope this work paves the way for a change in the current paradigm by developing understandable and human machine learning models to interpret reaction results that will increase creativity and chemical efficiency to find new chemical reactions and improve organic chemical pipelines and processes.”

Chopra stated that the machine learning approach that can be interpreted by the Purdue team, introduced as a chemical reactivity chart, can be extended to explore MCR reactivity or any chemical reaction. There is no need for large-scale robots because this method can be used by chemists when conducting reaction filtering in their laboratory.

“We provide the first report of a framework for combining rapid synthetic chemical experiments with quantum chemical calculations for understanding reaction mechanisms and statistically robust machine learning models that can be interpreted by humans to identify chemical patterns for predicting and experimentally testing heterogeneous reactivity. N-sulfonylimines, “in Chopra.

This work is in line with other innovations and research from the Chopra lab, whose team members work with the Purdue Foundation Research Office for Technology Commercialization for several technology patents.

“The unprecedented use of machine learning models in the production of chemical reactivity flow diagrams helps us to understand the traditional reactivity that is commonly used N-sulfonylimines in MCRs, “said Krupal Jethava, a fellow postdoctoral fellow in the Chopra laboratory, who authored the work. “We are confident that working directly with organic and computational chemists will open up new avenues for solving complex chemical reactivity problems for other reactions in the future.”

Chopra stated that Purdue researchers hope their work will pave the way for one example that will demonstrate the power of machine learning for the development of new synthetic methodologies for drug design and beyond in the future.

“In this work, we strive to ensure that our machine learning model can be easily understood by chemists who are not experts in this field,” said Jonathan Fine, a former Purdue graduate student who authored the work. “We believe this model has the ability to not only work to predict reactions but also to better understand when a reaction will occur. To demonstrate this, we use our model to guide additional substrates to test whether the reaction will occur.”

References: “Accelerated Reactivity Mechanisms and Machine Learning Models Translated from N-Sulfonylimines to Rapid Multicomponent Reactions” by Krupal P. Jethava, Jonathan Fine, Yingqi Chen, Sunday Hossain and Gaurav Chopra, October 19, 2020, Organic Letters,
DOI: 10.1021 / acs.orglett.0c03083

About Purdue Research Foundation Technology Commercialization Office

Purdue Research Foundation Office of Technology Commercialization Office operates one of the most comprehensive technology transfer programs among leading research institutes in the U.S. Services provided by this office in support of Purdue University economic development initiatives and benefits college activities through Purd commercialization, licensing and protection intellectual. The office recently moved to the Convergence Center for Innovation and Collaboration in the Discovery Park District, close to the Purdue campus. In fiscal year 2020, the office reported 148 deals completed with 225 signed technologies, 408 disclosures received and 180 issued U.S. patents. The office is managed by the Purdue Research Foundation, which received the Innovation and Economics of the University of Welfare 2019 for Venue from the Public University Association and Grants. In 2020, IPWatchdog Institute ranked Purdue third nationally in creating startups and in the top 20 for patents. The Purdue Research Foundation is a private, non-profit foundation created to advance the mission of Purdue University.

About Purdue University

Purdue University is a public research institute that is developing practical solutions to today’s most difficult challenges. The No. 5 Most Innovative University Rankings in the United States by US News & World Report, Purdue delivers world-changing research and discovery abroad. Committed to live and online, learning the real world, Purdue offers a transformative education for all. Committed to the ability and access, Purdue has a frozen tuition and high cost at the 2012-13 level, allowing more students than ever to pass debt free.

Related articles

Comments

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Share article

Latest articles

The origin of sun-shaped black holes and their relationship to dark matter

What is the origin of black holes and how does that question relate to another mystery, what is the matter of darkness? Dark...

Infectious diseases և social distance in nature

Mandrils take care of each other. This type of monkey continues to care for sick family members while actively avoiding sick individuals who...

Laser pulses act as an industrial-grade machining tool

Fused silica oil with a thickness of 350 um, selected by laser engraving with high quality glass lasers with diameters between 0.25 mm and...

Innovative and Non -Destructive Strategies for Analyzing Materials Returned From Mars

The distribution images of certain elements of one of the meteorites were analyzed by the group. Credit: IBeA / UPV / EHU The IBeA...

The new data transfer system is 10 times faster than USB and uses polymer cables as thin strands of hair

Researchers have developed a data transfer system that pairs high-frequency silicon chips with a polymer cable as thin curly hair. Credit: Offered by...

Newsletter

Subscribe to stay updated.