Automated Chemistry to Provide Tuberculosis Medicine to Low- and Middle-Income Countries

Through a merger of artificial intelligence and chemical synthesis, the laboratories of Drs. Tim Cernak and Paul Zimmerman at the University of Michigan (UM) will partner to lower the cost of and increase access to critical tuberculosis medicines.

According to the World Health Organization (WHO), tuberculosis, or TB, is a preventable and treatable bacterial infection spread through the air that usually affects the lungs but can also affect other areas of the body. Despite its treatability, the WHO estimates that 1.6 million people died from TB in 2021, making it the 13^th most common cause of death and the second biggest infectious killer after COVID-19. So why is a treatable disease so deadly? The answer is simple: not everyone can access the drugs used to treat TB.

“Modern tuberculosis medicines have increasing efficacy but remain challenging to access in developing countries,” said Cernak, Assistant Professor of Medicinal Chemistry at the UM College of Pharmacy. “Unfortunately, many modern TB drugs are very difficult to produce, synthesized by chemical reactions that are non-selective and low yielding.” This difficulty of production drives up the cost of TB drugs, which restricts developing countries’ access to life-saving therapies.     .

To solve this problem, Drs. Cernak and Zimmerman, along with their labs have teamed up to find a solution using artificial intelligence (AI). Driving up reaction performance using AI can help in driving down costs as this is essentially an optimization problem.  Analyzing and improving the reaction performance of compounds is essentially an optimization problem, and machine learning can speed that process. Previously these labs have shown that active transfer learning, a type of AI, can be used to predict the outcome of various chemical reactions. This approach relies on coupling high-throughput chemical experimentation run on robotic platforms in the Cernak lab with high-powered computational methods developed in the Zimmerman lab. Using the model, a feedback loop is set up and then the AI can learn from the experimental data generated in real-time. The labs then plan to use these tools to drive up reaction selectivity and drive down costs on a key TB medicine.  “This is where AI can shine,” said Cernak.       Drs. Cernak and Zimmerman will engage with several of their lab members in the effort, including students Andrew McGrath, Ruheng Zhao, Sam Zhang, Di Wang, Dr. James Douthwaite, Dr. Chris Audu, Bo Mahjour, Andrew Outlaw, Eunjae Shim, Dr. Yingfu Lin, Yu-Ting Kao, Dr. Haiyan Huang, Jordan Bench, Tesko Chaganti, Niharica Kannan, Jillian Hoffstadt, Dr. Jayabrata Das, Dr. Clinton Regan, Dr. Hamid Nodeh, Cameron Hempton, Dr. Xueying Zhang, Chun-Yi Tsai, Yannik Esser, Dr. Ying Tan, Dr. Qiyuan Zhao, and Dr. Nuwan Pannilawitha in the Cernak lab. Participants in the Zimmerman lab include Dr. Hosh Kammeraad, Dr. Soumi Tribedi, Dr. Taveechai Wititsuwannakul, Dr. Qiyan Zhao, Timothy Jugovic, Nicole Woodall, Exequiel Punzalan, Khoi Dang, Eunjae Shim, Jefferey Hatch, Kevin Rivera Cruz, Oleksii Zhelavskyi, Soumik Das, Nicole Orwat, Alex Stark, Scarlet Aguilar Martinez, Vaibhav Khanna, and Elizabeth Doty.

This innovative work. is funded through the Bill & Melinda Gates Foundation “We are humbled by this opportunity to apply our tech to an urgent medical issue,” said Cernak.

To follow the progress of this project, follow the Zimmerman lab on Twitter @ZimmermanUMich and the Cernak lab by following Tim Cernak on LinkedIn.