The projected US$2.6-billion cost of creating a therapy hangs over scientists looking for new treatments like a huge number. Since it includes money spent on the nine out of ten potential cures that fall short between phase I trials and regulatory approval, a large portion of that money effectively goes down the drain. Few experts in the field disagree that things need to be done differently.
A subfield of computer science known as artificial intelligence (AI) aims to replicate how the human brain makes decisions and solves problems. The application of AI is not new in the history of pharmaceutical discovery; it has been around for almost a century.
Despite not receiving much attention, artificial intelligence has been employed to promote the expansion of drug discovery’s sophistication over the years. Using AI models to assist in establishing connections between the structural characteristics of chemical compounds and biological action is a prime example.
They are crucial for drug discovery for Advanced Therapies and aid researchers in making more accurate predictions about how a drug candidate will behave in the body. Even while their predictions are constrained by the limitations of the models, they have greatly accelerated the process of discovering new treatments by enabling researchers to concentrate on compounds that have a better probability of combating specific diseases.
Leading biopharmaceutical firms think a remedy is just around the corner. In order to find immuno-oncology treatments, Pfizer is using IBM Watson, a machine learning system.
Sanofi has agreed to use the artificial intelligence (AI) platform of UK startup Exscientia to look for treatments for metabolic diseases, and Roche subsidiary Genentech is using an AI system from Cambridge, Massachusetts-based GNS Healthcare to support the global company’s search for cancer therapies. The majority of sizable biopharma businesses have comparable internal initiatives or alliances.
Small-molecule drug discovery can benefit from AI in four different ways: access to new biology, improved or unique chemistry, higher success rates, and speedier and less expensive discovery procedures.
The technique can solve numerous problems and limitations in conventional research and development. Each application gives drug discovery teams new information, and in certain situations, it might completely change tried-and-true methods. Understanding and differentiating between use cases is essential because these technologies are applicable to a range of discovery scenarios and biological targets.
We can now create a whole virtual environment for drug development, complete with in-silico models that mimic human disease and make use of enormous volumes of genetic, phenotypic, and chemical information. Access to and analysis of that data are both free.
We can detect illness traits that discovery approaches frequently miss because of their reliance on a single predetermined hypothesis by using computational methodologies and algorithms. Multiple targets can be tested simultaneously by possible therapies. AI bridges the gap for us, but it still requires our direction.
By applying AI to shorten the process of starting in vivo testing, we can accelerate the rate at which medications are introduced into preclinical research. With lightening speed, we can cross-reference libraries of candidate drugs against disease targets.
The viability of these chemicals in comparison to safety and efficacy markers can now be predicted more accurately. This level of development would require years to complete using conventional techniques, but by incorporating technology, we can complete it in a matter of weeks with only a few chemicals.
