The introduction of in silico modelling was revolutionary for drug discovery and biomedical research. Using computer simulations to conduct experiments, researchers can dramatically reduce the time needed to develop treatments. By enabling rapid testing and optimisation of drug candidates, efficiently screening extensive chemical libraries and modelling complex biological interactions, computer simulations reduce the need for lengthy and costly laboratory work.
During the COVID-19 pandemic, researchers used in silico models to investigate how various vaccines fought the virus. The results were remarkably consistent with those from human trials.
However, this computer modelling method is on the cusp of transformation itself, powered by generative artificial intelligence (AI).
With the ability to generate synthetic data, accelerate the process further and support treatment personalisation, the impact is likely to be vast. So much so, regulatory bodies such as the Food and Drug Administration in the US are already considering ways to manage this technology that balance safety and the potential for innovation.
In this article, we’ll explore how generative AI unlocks new biomedical research and drug development opportunities.
The role of AI within in silico drug design
-
1. Increasing accuracy with synthetic data
Traditionally, in silico modelling has relied on a combination of preclinical and clinical data and vast databases of chemical structures, DNA sequencing and pathways. When combined, in silico methods can analyse this information to predict drug behaviour and efficiency.
However, while helpful, the output is limited by the quality and quantity of available data, which is primarily problematic when there are data and knowledge gaps. For example, during the COVID-19 pandemic, there was initially limited understanding of how the virus interacted with the immune system, leaving a gap in the necessary data. Rare and complex diseases also present issues, as in silico modelling can only review available data – limited data means limited output.
By integrating generative AI and advanced machine learning (ML) to create synthetic data, researchers can begin to overcome some of these challenges. AI can generate datasets that closely mimic real-world data, filling in the gaps as needed.
Using this data, in silico models can better predict drug efficacy, side effects and how drugs may interact with human immune systems.
-
2. Accelerating drug discovery
On average, developing a new medicine takes 10-15 years and a cost of USD $2.6 billion. Of course, there are times when this process is accelerated due to global collaboration, increased funding and emergency conditions, such as the need to develop a vaccine during the COVID-19 pandemic.
However, by applying generative AI to in silico methods, drug development times may be significantly reduced. Companies leading the way with AI have reported the discovery of a novel drug target and novel molecule – a process that took less than 18 months and only 10% of the costs of a traditional project.
This acceleration was due to the rapid analysis and automation possible thanks to generative AI. The technology can quickly analyse vast datasets, patents and literature, identifying targets and candidates faster than traditional methods.
In addition, it can automate the research and development tasks, reducing time and increasing the efficiency of the process. This leaves the people involved with more time to apply their expertise to higher-value tasks.
-
3. Personalising treatments with generative AI
While it is possible to identify personalised treatment options using traditional in silico methods, the potential to scale this is limited due to the amount of data needed.
AI can create detailed simulations of individual patient responses to different treatments, predicting how each person might respond based on their genetic, physical and clinical data.
Without traditional limitations, pharmaceutical and biomedical research organisations could scale this, allowing for broader application of personalised medicine. In turn, this empowers people with more effective treatment options – and faster.
A data-first approach to drug discovery
Any form of AI relies on high-quality, comprehensive data. Without it, researchers cannot count on the technology to provide accurate and reliable insights or fully capitalise on the opportunity to accelerate drug discovery and development.
However, it’s also critical to maintain this data-first approach while using AI technology to support such complex and sensitive tasks. Drug development is an area of high regulatory standards, and the process must remain transparent and compliant.
With many AI tools, the outputs seem correct, but it’s often unclear to see how the technology reached its conclusion. In regulated industries, it’s critical to understand how these decisions were made.
By using AI tools designed with transparency at the core, pharmaceutical and biomedical research organisations can ensure governance and oversight throughout the process. It’s critical to choose technology that ensures each AI-taken action is logged.
Agentic AI for transparency and compliance
Our industry accelerator uses multi-agent autonomous teams to speed up complex workflows while offering full transparency.
In this technology, lead AI agents assign tasks to other AI agents. Each completes tasks and reports back to the lead agent, responding to feedback and adjusting course where needed. Critically, every step and action is logged, allowing for complete human oversight and compliance.
With this data-first approach and rigorous approach to transparency, even the most stringent requirements can be met, empowering researchers with the tools to confidently accelerate drug development.
To learn more about how generative AI can impact the healthcare and life sciences industry, download our whitepaper: Unlocking the Potential of Generative AI in Healthcare and Life Sciences