How is AI Transforming Drug Discovery?

Artificial intelligence (AI) is currently at the forefront of everyone’s minds, with its exciting potential to revolutionize every industry and alter traditional workflow. The growing impact of AI on many industries also includes the pharmaceutical sector, which may have major implications on the drug discovery process, making this otherwise lengthy process faster, more efficient, and cost-effective.1

AI in drug discovery: An overview

Conventional drug delivery is known for being both a lengthy and expensive process, with pre-clinical testing taking between three to six years and costing between hundreds of millions of dollars to even billions of dollars.1

Overall, 6-7% of the global gross domestic product, comprising approximately 8.5 to 9 trillion dollars, is used every year in the healthcare industry, with the cost of novel medicines being brought to the market being over 1 billion dollars and taking up to 14 years.2

The use of innovative AI tools has the potential to revolutionize this age-old process at almost every stage, including (i) target identification, (ii) molecular simulations, (iii) prediction of drug properties, (iv) de novo drug design, (v) candidate drug prioritization, and (vi) synthesis pathway generation.1

Accelerating research and development

During the target identification phase, AI tools can be used for large datasets, such as omics datasets and disease associations. This enables better comprehension of mechanisms underlying diseases and identifies potentially novel proteins or genes that can be targeted for innovative treatment.1

Currently, only approximately 3,000 proteins have been identified as potential therapeutic targets from the estimated total of 20,000 proteins in the human proteome. Future knowledge of the use of AI has the capacity to lead to a further understanding of which drugs may be therapeutic targets.2

The use of AI for predicting three-dimensional structures of targets may even be revolutionary for drug design, as when combined with other systems, such as AlphaFold, it can accelerate drug design to ensure effective binding to the target.1

An example of AI being used for drug discovery includes a deep learning algorithm that has been recently trained on a dataset of known drug compounds and their associated properties to suggest novel therapeutic molecules that have desirable characteristics. These suggestions can inform the fast and efficient design of novel drug candidates.3

Enhancing predictive accuracy

Significantly, AI can also be used to predict drug properties, with these tools being used to predict key properties of drug candidates, such as toxicity, activity within the body, and physicochemical properties. This can lead to a more optimized process, with a higher likelihood of the drug candidates being safe and effective for human use.

Drug-drug interactions occur when drugs are combined for the same or different diseases in the same patient, resulting in adverse effects, which can be problematic in the drug discovery process. The use of machine learning aims to address this problem by accurately predicting the interactions of novel pairs of drugs to reduce the risk of adverse reactions, accelerating the drug discovery process to develop more effective and safer medications.

An example of a successful application of AI in drug discovery includes the identification of novel compounds for cancer treatment, where researchers trained a deep learning algorithm on a large dataset of known cancer-related compounds as well as their associated biological activity.3 This research has significant implications for the future of cancer treatment, with applications in discovering novel drug candidates.3

Reducing costs and time

With the high expenditure and lengthy timeframe associated with the drug discovery and development process, the use of AI may be revolutionary in reducing these obstacles.1,2,3

AI has streamlined many stages of drug development, from synthesis to testing, with these innovative tools enabling researchers to focus on drug candidates with more promising efficacy and reduced toxicity. This can lead to an optimized drug discovery process, decreasing the time and cost of pre-clinical testing, with the AI tools also analyzing the large datasets, which can also influence drug design to match the therapeutic target effectively.1,2,3

In silico target fishing technology is an example of an AI tool that is used in the pharmaceutical industry to predict biological targets based on the chemical structure. Target fishing technology can be used to accelerate the process of selecting and identifying target proteins, which aids in decreasing the total experimental cost during drug development.4

Challenges and limitations

Although the benefits of AI have become popular, there are many challenges and limitations associated with using these innovative tools that require consideration.3

An important challenge faced by AI includes the availability of suitable data, with AI tools requiring a large amount of information in order to train the tool. Accessing the amount of data required may be limited, of low quality, or inconsistent, and this can ultimately impact the reliability and accuracy of the results.3

How AI Could Transform Drug Development And The Life Sciences
Currently, AI-approaches cannot be substituted for conventional experimental methods as they are not able to replace the expertise and experience of human researchers. AI-approaches can only provide predictions dependent on available data, however, the results require validation and interpretation by researchers.3

Combining the predictive abilities of AI with the expertise and experience of researchers, it is possible to optimize the drug discovery process, as well as to accelerate the development of novel drugs.3

Ethical considerations

Ethical considerations are a significant challenge with AI-approaches raising concerns about fairness and bias.

A key concern of the use of AI includes its decision-making capacity that can potentially impact the health and wellbeing of people, including which drugs to develop, which clinical trials to undergo, as well as how to carry out market distribution.3

The potential for bias within AI algorithms may cause unequal access to medical treatment and unfair treatment of various groups of people, which could undermine principles of both equality and justice.3

Job losses as a result of automation are also an ethical concern, with the progress of AI having a potential impact on workers, which may require policies to provide support for those who may be affected.3

Ongoing discussions and guidelines that aim to address these concerns include regularly reviewing and auditing AI systems and models for bias and having strong data privacy and security protocols.3

Future prospects and conclusion

AI is transformative for many fields, including drug discovery, with innovative models being used to progress personalized medicine and targeted therapies with an optimized drug discovery process and novel therapeutic targets.2,3

The impact of AI on the drug discovery and drug development process may be revolutionary, with a transformative potential to enhance every part of the process, from target identification to the effective design of drugs.2,3 With the continued innovation of AI models and software, this useful tool is ever-evolving and may have significant implications for the future of drug discovery.3

References

1. Chun M. How artificial intelligence is revolutionizing drug discovery – bill of Health. Bill of Health – The blog of the Petrie-Flom Center at Harvard Law School. March 8, 2023. Accessed May 23, 2024. https://blog.petrieflom.law.harvard.edu/2023/03/20/how-artificial-intelligence-is-revolutionizing-drug-discovery/.
2. Qureshi R, Irfan M, Gondal TM, et al. AI in drug discovery and its clinical relevance. Heliyon. 2023;9(7). doi:10.1016/j.heliyon.2023.e17575
3. Blanco-González A, Cabezón A, Seco-González A, et al. The role of AI in drug discovery: Challenges, opportunities, and strategies. Pharmaceuticals. 2023;16(6):891. doi:10.3390/ph16060891
4. Vora LK, Gholap AD, Jetha K, Thakur RR, Solanki HK, Chavda VP. Artificial Intelligence in pharmaceutical technology and Drug Delivery Design. Pharmaceutics. 2023;15(7):1916. doi:10.3390/pharmaceutics15071916