Active pharmaceutical ingredient (API) synthesis involves creating the substances that cause the desired, direct effects. API synthesis could concern one pharmaceutical compound or a mixture of several. They can be found in numerous forms, from powders to crystals.

This process requires expert input because it usually needs people to do multiple, complex steps. Moreover, synthesizing the API is essential in getting drugs ready for the market in formats patients can tolerate and work with maximum effectiveness.

Here are a few things people should expect to become more prominent among professionals and companies involved in API synthesis.

Artificial intelligence will improve API synthesis

People are increasingly interested in how artificial intelligence (AI) might make API synthesis more efficient. Early work in that area certainly shows potential. For example, scientists can save substantial time using machine learning algorithms that identify molecules with disease-fighting potential or other desirable properties. However, a potential downside is they identify candidates that may be too labor-intensive or impossible to produce in a laboratory at scale.

However, researchers at MIT developed a new approach that restricts AI algorithms so they only suggest synthesizable molecular structures. It solely provides options for molecules made from materials scientists could purchase and ensures that any reactions between them follow chemical laws.

This algorithm also works quickly, taking only one second to suggest synthetic pathways. Other AI solutions work similarly but have separate processes for finding molecules and evaluating their synthesis potential. This method takes minutes to complete. That still seems fast, but it can become inefficient as the number of potential molecules to screen rises.

In another example, a team at ETH Zurich developed an algorithm that helped them find molecules that behaved identically to natural substances but were comparably easier to make. Conversely, this AI application highlighted natural substances which could have useful pharmaceutical properties.

People familiar with API synthesis widely use certain compounds in their processes. Dimethyl sulfoxide (DMSO) is a common choice that also serves as an excipient. A sterile 50% aqueous DMSO solution is also approved to treat interstitial cystitis. AI could help uncover additional options, unlocking new potential for drug discovery and development and, eventually, treatment.

Collaboration will enhance API efforts

Keeping a pharmaceutical business running smoothly requires working with multiple parties. Some experts recommend a parallel supply chain approach whereby a company produces a small percentage of critical drugs locally, with the rest made overseas. Having numerous supply chain partners in various places is a best practice for avoiding slowdowns or difficulty obtaining essential products.

People looking to the future of API synthesis should expect more collaboration, too. One recent example involved four companies teaming up to provide a standalone service to build rapid-response mobile units for API production. The Production Intelligente de Principes Actifs (PIPAc) project includes professionals from NovAliX, Alysophil, De Dietrich Process Systems and Bruker.

One of the main goals of PIPAc is to sever the long and often complicated supply chains associated with API production. This initiative will handle synthesis, as well as continuous flow chemistry and in-flow analysis. It will also harness the power of AI to design optimized production units with autonomous capabilities.

PIPAc has an investment exceeding 3.5 million euros and a task force of employees from each of the four participating companies. There are also 50 additional scientists from elsewhere giving their input. Most people working on PIPAc are in France, with the rest residing elsewhere in Europe.

The collaborators hope to have their first demonstration of an industrial-scale mobile production unit ready by 2024. They believe it will increase the flexibility and resilience of API creation, making the solution able to meet the demands of today’s market. This is a good example of how people working together could improve API synthesis’s future.

Robots will support the synthesizing of APIs

People are also eager to explore what role robotics might play in API synthesis. Researchers from the U.S. Department of Energy’s Lawrence Berkeley National Laboratory and the University of Massachusetts Amherst built the first aqueous robot that powers itself and runs continuously without electricity. They believe it could assist in synthesizing chemicals automatically or get deployed for drug-delivery applications.

The scientists clarified that they use chemistry to control a robot’s buoyancy. Salt makes them have more or less density than the liquid surrounding them. Additionally, the machine gets its power from the surrounding media. The robots also transfer chemicals back and forth without stopping as long as they can get energy from their surroundings.

In another instance, the RoboRXN robot developed by IBM’s researchers can automate chemical synthesis by mixing molecules. More than 29,000 people have used the machine so far, generating approximately 5 million predictions about chemical reactions. IBM’s representatives say the system is more than 90% accurate.

The autonomous machine allows scientists to make requests remotely. One did so despite being 75 miles away from the robot. It also uses AI to determine the best combinations. Technological advancements like this one are why some people see scientists increasingly removed from the lab or at least engaging in different tasks than they once did.

The future of API synthesis looks bright

Trends like those described here illustrate why there’s a lot to look forward to and feel hopeful about regarding how APIs get synthesized. Some of the research efforts mentioned may not become commercially viable, but they’re still examples of progress in production methods and the industry at large.