When we are experiencing minor problems – such as an ache or pain – or have a more serious illness, most of us will rely on medicines to help us get better. But how many of us are aware of the huge amount of research that goes into the drugs we take?

Developing a new drug is a long and expensive process. It costs an estimated $2.6 billion and takes up to 12 years to take a drug through to regulatory approval [1].

The overwhelming majority of today’s pharmaceuticals are ‘small molecule’ compounds, composed of a few hundred atoms or less. Between 2010 and 2017, 76% of new drugs approved by the US Food and Drug Administration (FDA) were small molecules [2]. Due to their extra-small size, these drugs are more likely to get inside cells to reach their targets in the body compared to larger biological therapies.

This unimaginably diverse group of compounds share little in common except for their size and the fact they’re made from synthetic chemical reactions. A prominent example is aspirin, which is one of the oldest and most used medicines in the world – and many of the newest, most cutting-edge drugs for treating illnesses such as cancer, autoimmune diseases and depression, are also small molecules. And there are currently many more being developed to prevent and treat different conditions.

At the heart of small molecule drug discovery is chemical synthesis, which involves medicinal chemists creating brand new molecules through a complex, step-by-step process. Despite decades of research this is still a long, laborious procedure – and is a key bottleneck for advancing new medicines to the clinic.

But recent advances in AI-based software offer unprecedented new opportunities to help speed up this stage of drug discovery and get effective drugs to patients, faster.

Developing drug molecules is a slow, iterative process that involves sifting through thousands of candidates to find the most suitable one. Researchers need to carry out huge numbers of experiments across several rounds of optimization.

The process begins with biologists identifying a promising new drug target – such as an enzyme involved in a disease-critical pathway. The search then begins for potential drug candidates that can specifically interact with the desired target to cause a beneficial change. But to become a successful drug, a molecule will also need a host of other properties, including non-toxicity, solubility and stability.

Researchers will typically kick off by screening libraries containing many thousands of compounds, with the help of high-throughput assays and computer software. The aim is to narrow down several initial ‘hit’ compounds with activity against the target to move onto the next stage.

Next, the process of lead optimization begins, which involves medicinal chemists tinkering with the core chemical structure of a hit compound to fine-tune its properties. This may involve taking out parts of a molecule or adding in others that are predicted to improve its characteristics, which requires a thorough understanding of the chemical reactivity of complex molecules.

Each hit will then undergo a battery of tests to explore its suitability as a drug. Finally, only a single compound with the best qualities will remain – the drug candidate. This is followed by steps to ensure the feasibility of its large-scale synthesis and manufacturing. If it passes, the drug is ready for testing in pre-clinical studies– which will hopefully follow on to clinical trials in people.

But unfortunately, most compounds will never reach the clinic, with vast numbers discarded at some stage along the way.

Synthetic chemistry has largely relied on the knowledge and expertise of the medicinal chemists and their ability to come up with original ideas for synthesizing new molecules. But recent advances in AI may help boost their chances of success, speeding up the progress of drug discovery projects.

“We have acquired and continue to develop a new tool which can save chemists a lot of time and their companies a lot of money and help ensure that drugs get to patients faster,” says Trice.

Engineered by chemists and computer scientists for more than 15 years, SYNTHIA™ retrosynthesis software is powered by sophisticated algorithms that can help experts’ access and make use of the vast amounts of data on chemical synthesis collated over decades of research.

“Our new drug discovery software provides chemists with many different routes to attempt to make their molecules,” explains Trice. “It explores new and known solutions, eliminates options that won’t work, and narrows down the most promising pathways to explore.”

The tool works by harnessing the potential of advanced algorithms powered by more than 100,000 hand-coded reaction rules – painstakingly sifting through retrosynthetic possibilities while at the same time examining what has been done, what could be done, and what starting materials are available.

“We write the rules with expert organic chemists, which makes our tool entirely different from others out there.”

SYNTHIA™ retrosynthesis software provides invaluable information that can point towards the best possible route to execute – minimizing cost, the number of steps, and the best chance of making the required molecule with the desired properties. This can dramatically reduce the time it takes for a chemist to think of a viable route to embark on in the lab.

“It shortens the time of ideation to actual execution in the lab,” Trice continues. “More broadly, software like this can shorten one of the longest steps in the drug discovery process.”

Most small molecule drugs are the end-product of painstaking laboratory work – the result of an intricate process of hypothesis, design, and synthesis of individual molecules and testing their effects in biological systems.

We hope that SYNTHIA™ retrosynthesis software can significantly speed up this process, accelerating drug discovery so that people can benefit from new life-changing medicines sooner. But it has the potential to do even more.

“A lot of groups are starting to look at the predictive piece,” says Trice. “While SYNTHIA™ retrosynthesis software and other tools can tell you to do Reaction A followed by Reaction B, the goal is to say while you are doing Reaction A, these are the exact conditions including temperature, reaction time, solvents and catalysts to make it successful.”

But the ultimate endorsement for our new tool should come directly from the experts who use it.

“Our first users were some of the most influential people in the chemistry community,” says Trice. “We hope that if they’re happy, they will tell their friends – helping to build momentum through ‘word of mouth’.”

To contact us, find out more or request a demo please click HERE

If you would like to get some more information about SYNTHIA™retrosynthesis software, you may be interested in reading our latest publication in the Science Mag.

What if you could use AI and machine learning to design plausible routes to complex natural products to find the best method for synthesizing these products? In Nature Research (Publishing), read how our SYNTHIA™ retrosynthesis software helps scientists reach their product goals, faster.

References:

[1] https://syrris.com/applications/drug-discovery-and-development/
[2] https://www.biopharmatrend.com/post/67-will-small-molecules-sustain-pharmaceutical-race-with-biologics/

Further information about SYNTHIA™ retrosynthesis software can be found here: https://www.sigmaaldrich.com/UK/en/services/software-and-digital-platforms/synthia-retrosynthesis-software