Pharma R&D Today
Ideas and Insight supporting all stages of Drug Discovery & Development
Can predictive retrosynthesis become a valuable part of a chemist’s toolkit?
Posted on January 18th, 2022 by Ann-Marie Roche in Chemistry
“Yes. I was amazed by the Reaxys Predictive Retrosynthesis technology,” says Moritz Classen of the Carreira Research Group at ETH Zurich. He and his colleagues spent months evaluating the tool as they applied it into their daily workflows. “It’s already very user-friendly and really does save a lot of time”. So, what can chemists exactly expect when using predictive retrosynthesis tools?
Moritz Classen champions the power of synthetic organic chemistry with infectious enthusiasm. “I look at chemistry as a sort of Lego set of the world. And by understanding how to combine chemical knowledge and chemical reactions, we can build fantastic things. We try to recreate what nature is doing but then in the lab. And meanwhile, the building blocks and how to connect them keeps evolving with every new discovery.”
Moritz is a PhD student at the Carreira Research Group at ETH Zurich. Not only is Professor Erick Carreira the editor-in-chief of the prestigious Journal of the American Chemical Society, he also rates as one of the most renowned organic chemists of his generation.
“It’s a very special research group for many reasons,” says Moritz. “One obvious reason is Professor Carreira being such a successful chemist. He is extremely knowledgeable and has built his reputation on great collaborative efforts. So, he attracts people who are not only very motivated but also very interested in working together.”
Much to learn
Moritz’s specialty is natural product synthesis. He’s currently working on a class of diterpenoids found in spurge plants, which have been used in traditional medicine around the world for thousands of years. “They are very bioactive and have already been applied to treatments with varying success. But there’s still much to learn.”
Hence, he is open to any technology that accelerates his research – particularly when it relates to his retrosynthesis execution. In an ideal R&D world, such a tool would continually deliver trustworthy and scientifically robust predictions in an easy and straightforward manner.
How does Reaxys Predictive Retrosynthesis contribute to execution?
Reaxys’ predictive module is fully integrated into its Retrosynthesis feature. This AI-fueled tool has been trained on 15 million unique single step reactions and 100 million virtual negative reactions. From this, roughly 400,000 reaction rules have been derived.
“You have it all in front of you,” enthuses Moritz. “Users can see the supporting literature for each predicted step in a route, along with commercial availability of starting materials – from well-known suppliers to smaller custom synthesis organizations.”
“Having this great access to all the literature is really beyond what even a very experienced chemist has. It can really help you decide whether a predicted route can work. And sometimes it also finds these sometimes very obscure disconnections and reactions, which can really help your lateral thinking,” says Moritz.
“We were obviously very happy to work with this groundbreaking technology. Those of us working with more complex molecules were less happy, because some steps or reactions require very specific reagents or conditions that might not be available. But I think that will come when the tool becomes more customizable and chemists can be more hands-on in guiding the process.”
“But in general, the tool is already very user-friendly and really does save a lot of time.”
Tips for adopting predictive retrosynthesis tools
Moritz came into the collaboration fresh. “Of course, I was familiar with Reaxys since I use it daily for literature, reagents and more. But I had no experience with predictive retrosynthesis tools.”
“I learned to shed some biases and be more open-minded. This isn’t something that’s out to replace you. At the same time, you shouldn’t assume it will be inferior to your own knowledge. Start with feeding the tool some molecules you are already familiar with and see what it comes up with. You’ll be surprised what it might find,” advises Moritz.
“I saw the information breakdown in three ways. The smallest proportion was the chemistry you already know works. Then came the chemistry you already know doesn’t work because you have tried it in the past. But the biggest proportion was the chemistry you didn’t know anything about – which can be super inspiring,” says Moritz.
“I think this tool can become a very important part of planning a synthesis. And it’s not as if a chemist can sit down for an hour and come up with a nice 20-step synthesis plan. There’s always a lot of trial and error. And a lot of times it can look good on paper, but then it turns into that Mike Tyson quote: ‘Everybody got a plan until they got punched in the face’. It’s our job to try it until it works. And this tool lets you stand up again faster.”
On the path to continual improvement
The Elsevier Reaxys team continuously looks for feedback on potential improvements from its users. And this collaboration with Moritz and the Carreira team was particularly inspiring in understanding the needs and priorities of the platform’s users.
“Already, Elsevier has applied more customization options which is great,” comments Moritz. “And I am really looking forward to continuing the collaboration. It’s truly fantastic to work with such a great team on such a great field that is developing so rapidly.”
Join the fun – and the learning curve. Watch the full Reaxys Retrosynthesis webinar with Moritz Classen to see the specific reaction routes tested by the Carreira Research Group.
R&D Solutions for Pharma & Life SciencesWe're happy to discuss your needs and show you how Elsevier's Solution can help.
Director, Corporate Markets Marketing, Elsevier
- Elsevier and LG: Turning data into action
- Pharma can now track the most relevant patent info – fast and at scale
- AI-driven innovation in life sciences: Unlocking the page – 20+ years of digital innovation at Elsevier
- AI-driven innovation in life sciences: Data as data should be
- AI-driven innovation in Life Sciences: Data Science & AI amplifies your innovative drug R&D capabilities