Drug discovery has traditionally been a costly and complex process fraught with significant challenges. An essential part of this process is understanding drug metabolism and pharmacokinetics, crucial determinants of a drug’s safety and effectiveness. These attributes are traditionally assessed using in vitro and in vivo studies, often leading to late-stage failures (Cook et al., 2014). However, the landscape of drug discovery is transforming thanks to the advent of artificial intelligence (AI), particularly generative AI, which promises to accelerate drug discovery and minimize late-stage failures.
Generative AI refers to machine learning models that generate new data instances resembling input data. In drug discovery, they can predict drug metabolism and pharmacokinetics, potentially accelerating the drug discovery process (Chen et al., 2018).
In the vast domain of hepatobiliary and metabolic disorders, bile acids (BAs) serve as crucial diagnostic biomarkers, offering a window into the underlying pathophysiology. The nexus between BAs and lipid as well as glucose metabolism is orchestrated via the farnesoid X receptor (FXR). Amidst this backdrop, the study ventured into unexplored territory to discern the impact of an exogenous FXR modulator, Ivermectin (IVM), on the plasma BA profiles in rats.
Understanding Drug Metabolism and Pharmacokinetics
Drug metabolism and pharmacokinetics (DMPK) studies are vital to understanding how a drug will behave in the human body. The information obtained from these studies is essential in determining the safety and efficacy of new drug candidates. Predicting these attributes accurately is often challenging, leading to delays, unexpected findings during clinical trials, and even drug withdrawals (Scannell et al., 2012).
Enter Generative AI
With the advent of AI, it is possible to make accurate predictions about drug metabolism and pharmacokinetics. AI algorithms, especially generative ones, can learn from existing data to predict the properties of new, unseen drug molecules. This approach can save time and resources on synthesizing and testing new compounds. The outcome is a more efficient, cost-effective, and faster drug discovery process (Stokes et al., 2020).
Generative AI in Predicting Drug Metabolism and Pharmacokinetics
Generative AI has the potential to predict how a drug will be metabolized in the body and its pharmacokinetics properties. With the help of large datasets and machine learning algorithms, we can train AI models to predict these properties with impressive accuracy. These predictions can guide scientists in the design of new drugs, reducing the risk of late-stage failures (Segler et al., 2018).
Real-world applications of AI in the field of Drug Metabolism and Pharmacokinetics (DMPK)
1. Insilico Medicine
Insilico Medicine is a biotechnology company that leverages generative AI to accelerate drug discovery. In a study published in Nature Biotechnology in 2019, the company described using their generative tensorial reinforcement learning (GENTRL) model to design, synthesize, and validate a novel drug candidate in just 46 days (Zhavoronkov et al., 2019). This is a significant improvement considering that traditional drug discovery processes can take several years.
2. BenevolentAI
BenevolentAI is another company applying AI to drug discovery. They use machine learning algorithms to understand the underlying mechanisms of disease, identify potential drug targets, and predict how drugs will behave in the body, including their metabolism and pharmacokinetics. The company has numerous ongoing projects, including collaborations with AstraZeneca and Novartis to leverage AI for improved drug discovery and development (BenevolentAI, 2021).
3. Atomwise
Atomwise is a startup that uses AI to expedite the drug discovery process. Their AI model, AtomNet, is capable of predicting the bioactivity of small molecules, aiding in understanding their pharmacokinetics properties and potential metabolism in the body. Atomwise has partnered with multiple pharmaceutical companies and academic institutions to discover drug candidates for various diseases, demonstrating the potential of AI in DMPK studies (Atomwise, 2021).
4. Deep Genomics
Deep Genomics uses AI for drug discovery, especially for genetic diseases. Their platform, Saturn, leverages machine learning to predict genetic alterations’ effects and design molecules to counteract these effects. The AI predicts the pharmacokinetic properties of these molecules, facilitating the selection of drug candidates (Deep Genomics, 2021).
In conclusion, these real-life examples demonstrate the potential of AI to accelerate the drug discovery process, particularly in the field of DMPK. As AI technology evolves, we can expect even more significant advancements in this area.
Looking to the Future
While generative AI offers a promising solution to some of the most significant challenges in drug discovery, it’s important to remember that these technologies are tools to aid scientists, not replace them. We still need skilled scientists to interpret the results generated by AI and apply them to the real world. As we continue to improve these technologies, they will become more accurate and valuable tools in the fight against disease (Cohen et al., 2020).
In conclusion, drug discovery advancements are increasingly driven by generative AI. By enabling more accurate predictions of drug metabolism and pharmacokinetics, these technologies can significantly accelerate drug development, reduce costs, and, most importantly, bring more effective treatments to patients faster.
References:
- Cook, D., Brown, D., Alexander, R., March, R., Morgan, P., Satterthwaite, G., & Pangalos, M. N. (2014). Lessons learned from the fate of AstraZeneca's drug pipeline: a five-dimensional framework. Nature reviews Drug discovery, 13(6), 419-431.
- Scannell, J. W., Blanckley, A., Boldon, H., & Warrington, B. (2012). Diagnosing the decline in pharmaceutical R&D efficiency. Nature reviews Drug discovery, 11(3), 191-200.
- Stokes, J. M., Yang, K., Swanson, K., Jin, W., Cubillos-Ruiz, A., Donghia, N. M., ... & Collins, J. J. (2020). A deep learning approach to antibiotic discovery. Cell, 180(4), 688-702.