How can scientists streamline drug discovery, a process that takes up to 15 years and costs around $2.6 billion? The convergence of biology, chemistry, and technology is changing the landscape for pharmaceutical companies. In the last decade, AI has stepped into the field, providing cost-efficiency and greater speed and precision to developing medicine.
Pharmaceutical companies are investing in AI, including Pfizer, one of the most prominent players during the outbreak of COVID-19. With multiple benefits in the biotech sector, the global use of AI in the drug discovery market is expected to account for USD 24,618.25 million by 2029.
AI can deliver value in drug discovery by identifying new therapeutic targets (biological molecules, typically proteins, associated with disease processes). This blog focuses on the role of virtual screening in drug discovery, a process powered by AI.
How do scientists find drug candidates?
The process of finding drug candidates consists of two main steps:
- Discover a therapeutic target: a biological molecule or a biomolecule that plays a critical role in the disease process at the cellular level.
- Identify molecules that might serve as drug candidates. Scientists achieve this through high-throughput screening (HTS), a process of automated testing of many chemical compounds to discern their biological activity and, thereby, a therapeutic potential based on the effect they exhibit in the test.
Even though the HTS technique has significantly improved, it can be costly and time-consuming. Traditionally, HTS has access to large and diverse chemical libraries; screening millions of molecules can take months.
Virtual screening in action
Scientists can streamline the HTS process with the use of virtual screening. This approach involves screening chemical compounds in silico utilizing computer simulation or modeling. It enables the identification of small molecules (ligands) that may bind to a therapeutic target and modulate its function, thus correcting its detrimental role in disease development.
For virtual screening to succeed, we must know the biological target’s three-dimensional (3D) structure in as much detail as possible. Recent advances in the AI tool AlphaFold, developed by DeepMind (a British subsidiary of Alphabet, Inc.), have enabled the prediction of the 3D structure of nearly all proteins found in living organisms with a high level of accuracy.
Dive deeper into AI and pharma
Virtual screening offers a fast and efficient way to identify new drug candidates and design new chemical compounds that could enable novel treatments. Learn more about the challenges and advances in virtual screening in our white paper.
