How AI and Machine Learning can change the lab forever

AI Machine Learning

The pharmaceutical industry is not exactly known as the most progressive sector when it relates to the technologies being employed. In many cases, the work in the laboratory is still being carried out exactly as it was in the 1940s. Starting in most cases a just a hypothesis, medicines are developed, tested & trialed, and, in the small case of success, further developed into a mass product at an enormous expense. The R&D costs per developed product in the pharmaceutical industry have been rising enormously for decades. As technology evolves, this is becoming unnecessary. Today much more is possible in the lab using AI and Machine Learning.

Numbers and figures

The figure below clearly shows the bottlenecks in the development of new medicines. Those are:

  1. The convoluted regulatory hoops that companies must go through before they can approve a product. This process can often take 12 years or more. In the meantime, you are also paying for staff, laboratories, and so on.
  2. The “outdated” method of working results in huge numbers of preciously developed medicines ending up lost. This problem tied in with the combination of the increased complexity of medicines over the years is making it harder and harder to develop medicines in the traditional way.
2-Crossing-over-the-Valley-of-Death-The-distinct-stages-of-drug-development-from.png
Source: Lost in Translation – Bridging the preclinical and clinical worlds concepts, Examples, Successes and Failures in Translational Medicine by Attilla Seyhan

An example to follow

That being said, the development of coronavirus vaccines is an enormous deviation from the norm. With the whole world working on this problem, working till late at night in labs, and removing as much red tape as possible, it is apparently possible to develop, approve and mass-produce vaccines within the span of a year. So, it can be done, but evidently only in special circumstances. For most drugs, approval is still a long and convoluted journey, and the “Valley of Death” needs to be crossed before any success looms on the horizon. Or, is there another way to approach this?

AI and Machine Learning

The term “industry 4.0” has become more common in our lexicon in recent years. It stands for the increasing possibilities to speed up production processes, make them more efficient and make them safer with the help of Artificial Intelligence and Machine Learning. You can read exactly how that works by checking out our special AI web topic.

Development on the computer

AI and Machine Learning also offer enormous potential to the pharmaceutical industry. By shifting a large part of the development process from the lab to the computer, development can be done faster and more efficiently. Data is stored centrally, allowing pharmaceutical companies to create an enormous pool of information that can be shared and accessed more easily.

Building on your predecessors

Due to these changes in the work process, it is possible to speed up the trajectories of new developments considerably. You can build directly on the knowledge of your predecessors, using their information as a springboard. This is a huge advantage, especially when it comes to complex products such as vaccines. By testing a new hypothesis on the computer and calculating an outcome based on Machine Learning and AI, you can significantly reduce the effort and manpower required to succeed in the lab.

A blueprint ready to use

An additional advantage is that in this way you create a blueprint in which all information is described exactly. Sensors also make it possible to link the information about the lab conditions in which the successful experiment was conducted to this. This means that almost all information is available that is necessary for further scaling-up and production. You also create a perfect log in this way that can be submitted to the controlling authorities in the later stages. Provided that all GMP requirements are met, you can considerably shorten the entire process in this way.