AI For Discovery Of New Drugs

      Comments Off on AI For Discovery Of New Drugs

AI For Discovery Of New Drugs

Leveraging the fashion effects surrounding artificial intelligence, many startups in the pharmaceutical industry claim to capitalize on profits. However, what real benefits can be expected from the application of artificial intelligence, and which applications can facilitate the search for new drugs? There are 1,060 molecules that could provide more potential drugs than all the atoms in the solar system, according to MIT. This offers virtually unlimited chemical possibilities.” To explore the sea of ​​molecules, researchers are turning to artificial intelligence (AI). Artificial intelligence (AI) selects molecules and associates them with potential targets such as proteins or cellular receptors. In July 2019, the Australian team designed an influenza vaccine using an algorithm designed adjuvant. And in February 2020, Insilico Medecine succeeded in developing a treatment for fibrosis in just 46 days thanks to AI.

Long and costly process…

The search for new drugs, especially small molecules (as opposed to biological molecules, larger, more complex and less stable molecules) can be viewed as a series of steps to identify and filter molecules. It has 4 stages.

First, the pathophysiological study of the disease allows you to identify the object of treatment (virus, enzyme, hormone, etc.). Next, hit generation creates the first list of molecules selected for activity on the target. Lead identification still allows you to isolate hits that best meet a limited set of biophysical, chemical and industrial criteria: the specificity of the target, the efficiency of the cell test, reproducibility and synthesis potential, patentability, etc. Continues.

Cause of decline…

The first reason is that pharmaceutical research deals with pathology with more complex physiology that requires therapeutics with more unique mechanisms of action. That said, diseases that are easy to treat already have cures and have given way to cancer, hereditary diseases and other complex pathologies, and researching and studying treatments requires very extensive knowledge and requires enormous resources to be mobilized. Another reason is technical bias. Without increasing the quality. Hence, more molecules have evolved further in the value chain with resources consumed to reach them without better leads. In short, a lot of money is spent on bad molecules due to heterogeneous industrialization of value chain processes.

Prediction, representation, exploration, generation…

The key to the efficiency of pharmaceutical R&D lies in part in these challenges. That is, maximizing very diverse criteria and finding molecules that will be tested step by step (i.e. better identification) over time. filter).

Artificial intelligence (AI) allows you to build a holistic model of drug design in which these problems are solved at the same time in all complexities from scratch.

Increased interest…

Artificial intelligence sells well and is poorly caught. And entrepreneurs understood it well. We have identified more than 100 companies (mainly startups) participating in the joint trajectory of pharmaceutical R&D and AI. There are already so few people who can actually offer it.

Towards a change of way…

AI will certainly contribute to creating a new identity for pharmaceutical research. Gradually scouring the budget of benchtop research and computer chemistry, laboratory investment in AI is becoming an increasingly important part of whether developing knowledge internally or externally using the know-how still weakly acquired in the industry.