Chemical.AI nurtures AI to become a true partner to chemists, says CEO Dr Ning Xia

1. Could you provide a brief overview of Chemical.AI and its solutions?

Chemical.AI is an innovative company founded in 2018, but we started the journey with chemoinformatics in 2008. We were one of the earliest players in this field to integrate AI with chemistry. We utilize proprietary deep learning algorithms, ever-growing chemical big data and state-of-the-art robotic lab automation to significantly improve the efficiency of chemical synthesis. Chemical.AI has a global presence with offices in Shanghai, Wuhan, Singapore, Boston, and Toronto and a team of 150 experts from various disciplines.

Our AI-driven solution includes four major products:

ChemAIRS is an AI-powered retrosynthesis routes design platform. It can design multiple synthetic routes for a target molecule within a few minutes, fully considering factors including reaction feasibility, functional group compatibility, chiral synthesis strategies, and many more key factors in synthesis, providing chemists diverse viable synthetic ideas.

ChemAIoT provides AI + automation solutions for chemical labs. With fully integrated software and hardware automation workflow, the lab supports 24/7 high throughput experiments and optimizations of conditions and processes.

ChemAIOS is an intelligent lab management system that helps with project and inventory management as well as data ecosystem building. It also has an electronic lab notebook (ELN) and quotation tools to enhance research and development efficiency.

ChemAILab provides chemistry CRO services by combining the capabilities of ChemAIRS, ChemAIoT and ChemAIOS. It serves various modalities including small molecules, ADC, natural products, target protein degradation, peptides and more. It provides not only customized synthesis but also scale-up synthesis from milligrams to kilograms.

2. How did the idea of combining chemical expertise with artificial Intelligence come about?

In the realm of chemistry, the single bonds and double bonds are akin to the binary digits of 0 and 1 in the world of computers. They possess the power to generate an array of creations through countless, seemingly monotonous combinations. My mother is a computer scientist so I encountered the beauty of data and algorithms at a rather young age.

After getting my Ph.D. in organic chemistry from Centre National de la Recherche Scientifique (CNRS), I actively seek opportunities to combine human chemical expertise with artificial intelligence to improve the efficiency of chemical R&D.

With countless possibilities in terms of combination, functional groups, and reaction conditions, designing synthetic routes traditionally demands years of training and trying. In drug discovery, only well-trained experts with over 10 years of experience are allowed to work on retrosynthesis route design. So I want to change this situation and it’s what I’ve been doing for almost 15 years now.

Now with our retrosynthesis planning platform, designing a plausible route takes about 15 seconds, while traditional approach typically takes 30-60 minutes.

3. How did the company foster innovation through the recent collaboration with Nanyang Technological University, Singapore?

In collaboration with NTU, we are poised to revolutionize the field of nanomaterial research and development. Through our partnership, we will assist NTU in consolidating experimental data, establishing comprehensive databases of reaction conditions and results, and developing AI models that optimize reaction conditions for nanomaterial synthesis and identify compound ingredients based on optical spectrum analysis.

By spearheading the development of NTU’s ground-breaking automated research platform for nanomaterials, we’re also gaining invaluable insights into the needs of chemical labs across the material industry. Chemistry, as a fundamental discipline, holds immense potential to drive transformation in various sectors.

Our advanced algorithms empower chemists in dry labs by swiftly generating retrosynthetic routes in a matter of minutes. Exhaustively exploring every possibility, it also meticulously ranks routes based on factors such as the number of steps, safety considerations, cost implications, and more.

In wet labs, we have developed a unique synergy between AI and automation, enabling our hybrid robots to conduct 24/7 high-throughput experiments. This seamless integration of AI and automation allows scientists to manage massive amounts of data efficiently, providing chemists with invaluable insights.

With automated workflows, our AI solutions enhances nanomaterial R&D processes, allowing scientists to focus on their core competencies and creativity.

4. What challenges do you foresee in integrating AI in the chemical sector, and how does Chemical.AI address these challenges?

When I first contemplated the fusion of AI and chemistry, one of the crucial questions was whether we should approach it as a data-driven endeavour, where AI simply acts as a massive calculator, or should we take a bolder path, feeding and training AI with data and information, moulding it into a true ‘chemist’? I decided to go with the latter.

AI + chemical synthesis is not just about data and calculation. It necessitates a unique blend of chemistry and computer science expertise. It entails a deep understanding of available raw materials, uncovering untapped potential and overlooked pathways. It requires the ability to ‘calculate’ synthesis strategies that yield optimal performance.

At Chemical.AI, our focus is on training our AI to grasp the underlying principles that drive its application. This includes understanding the intricate connections between proteins, mechanisms, and diseases, as well as identifying the specific features and functions required by novel materials. We are committed to nurturing AI to become a true partner to chemists, augmenting their capabilities and accelerating the pace of scientific discovery.

5. How does Chemical.AI differentiate itself from other players in the market?

I’m very proud of our self-developed AI algorithm. It represents the seamless integration of artificial intelligence and invaluable human experience. It designs routes based on over 60 million chemical data, including leading SCI journals and patents.

What sets our platform apart is that it was designed by experienced chemists who truly understand the real needs of a chemistry lab, so it’s been carefully designed to tackle real lab challenges from the very beginning. In a blind head-to-head comparison, ChemAIRS outperformed eight other prominent players in the field when they were challenged to propose synthetic routes for 60 structurally diverse and complex target molecules. The results spoke for themselves.

We’ve also been working on the automation of chemistry labs where we integrated our AI algorithm with robots to foster 24/7 high-throughput experiments. Our AI algorithm is compatible with a wide range of robotic arms, so we can not only provide a comprehensive AI + automation solution but also have the flexibility to integrate our solutions into existing software or hardware, offering a seamless upgrade path for those seeking an upgrade in their labs.

As we recently released a new feature called impurity prediction with our AI platform, scientists now can accurately predict impurities based on given molecular weights, significantly reducing the challenges associated with impurity analysis.

6. How does Chemical.AI envision its impact in the market in another five years, especially concerning advancements in AI and the evolving needs of the chemical industry?

In the next five years, I think AI-powered automation will have a much bigger impact on chemistry R&D. The market is poised for a revolution as the influence of AI + automation expands beyond industry giants, permeating middle-sized or even small companies and laboratories in public research centres. The potential applications of such AI-powered solutions are boundless, they will be tested and applied across many more industries to enhance research efficiency.

High-throughput experiments will be widely adopted in the discovery of medicines and novel materials to lower costs. By using AI and automation strategically, the speed of innovation will soar beyond human limitations. This trend is expected to continue as more chemistry data is generated.

7. Can you discuss ongoing research initiatives or future projects the company is particularly excited about?

At Chemical.AI, we are constantly exploring new frontiers and expanding the applications of our cutting-edge solutions. As I mentioned earlier, the realm of chemistry permeates countless industries, and we are passionate about making a positive impact across each of them.

Our track record in drug discovery is a testament to our capabilities. We have successfully collaborated with major pharmaceutical companies, assisting them in designing hit and lead compounds, optimizing synthetic routes, and even synthesizing new drug molecules. We have been also instrumental in helping these companies establish automated labs, revolutionizing their research and development processes.

Our technology has also been well utilized in the fields of materials and chemical engineering. By customizing automation labs tailored to their specific procedures, we empower researchers and engineers to achieve remarkable advancements in their respective domains.

The possibilities are limitless, and we are excited to continue forging new paths, embracing novel challenges, and making a lasting impact across diverse industries.

8. How has the journey with Chemical.AI impacted you personally? What connects you the most with what you do?

I am very lucky to be doing what I have always dreamed of doing. My lifelong passion for programming and chemistry began during my college years when I found the two disciplines so fascinating that I began planning to merge the two ideas in my future work.

Chemical.AI was born as the perfect fusion of these two beloved fields, combining my love for programming and chemistry into one cohesive entity. Personally, I have enjoyed every moment at Chemical.AI as we continue to push the boundaries and shape the future of chemistry through the power of artificial intelligence.

The most exciting thing would be finding how a beautiful chemistry process can be coded and exploring it from a fresh perspective. We strive to build a bridge between real-world chemical reactions and the digital world, unlocking countless new discoveries that have remained hidden until now.

What truly motivates me is the impact our work can have on people’s quality lives. We understand the challenges in drug discovery, including the exorbitant costs. By harnessing new technologies, we aim to help reduce these expenses significantly so the development of life-saving drugs could be more feasible, even for rare diseases. This is what inspires my team and me to work diligently and provide our state-of-the-art solutions to more companies, ultimately benefiting as many people as possible.

To read more such interview, head to the link here.