An expert interview with Stephan Hegge, PhD, VP of Corporate Strategy at HotSpot Therapeutics and Thomas Wollmann, PhD, CTO at Merantix Momentum on AI in identifying drug targets and unmet medical needs.

Vendela: Thank you both for being here today. Let's get started with the interview. Stephan, could you briefly introduce yourself and talk about the complexities and challenges of identifying unmet medical needs?

Stephan: Certainly. I'm Stephan Hegge, VP Corporate Strategy at HotSpot Therapeutics. HotSpot is a Boston-based biotech company focused on new drug development. Identifying unmet medical needs requires looking at the patient population and understanding their treatment options. We aim to fill therapeutic gaps where current drugs are not effective, safe or practical. Our company specializes in finding allosteric pockets on proteins using a unique approach. This process uses platforms that rely on Big Data and occasionally integrate machine learning or AI. After we generate a list of potential targets, experts in specific therapeutic areas narrow the list down to a few proteins of interest. We then perform detailed analysis to assess the diseases, patient populations, and unmet medical needs associated with these proteins of interest. We are developing small molecules that aim to manipulate our target of interest in a way that we believe will translate into a clinical effect that addresses these unmet medical needs.

Vendela: Let's talk about the impact of AI on the drug development process. How has AI changed the way you manage processes and challenges in your team?

Stephan: AI is certainly having an impact on the pharmaceutical industry, including HotSpot. Different companies have different approaches to AI, some aiming to eliminate human bias by leveraging all available data. In contrast, our philosophy is that human expertise is second to none when it comes to innovation and outside the box ideas. Therefore, we use AI primarily to scale human expertise rather than replace it. Once we understand where our expertise comes from, we can train our machines. For example, we have applied this approach in the context of natural language processing to identify and analyze publications that describe potentially relevant targets for us. Generally speaking, the field continues to evolve, and companies are exploring different philosophies and applications for machine learning.

Vendela: Thank you, Stephan. Thomas, as CTO of Merantex Momentum, what challenges do you see in the pharmaceutical industry, especially in manual data analysis and demand generation?

Thomas: In the pharmaceutical industry, one of the main challenges is filling the pipeline with potential drug candidates. This requires screening different compounds and understanding the unmet needs they could address. Data for this analysis comes from a variety of sources, including in-house data, acquired data, academia, and social media listening. Proper data integration to identify the right candidates is time consuming. Automation is challenging due to gaps where human intervention is required. AI can play a critical role in exploring and finding valid candidates, improving the efficiency of data analysis in pharmaceutical research.

Vendela: Could you explain how AI algorithms effectively process structured and unstructured data and their relevance in pharmaceutical research?

Thomas: AI algorithms, especially machine learning, are great at finding patterns in data. This can be done in an unsupervised way to identify data clusters or patterns, or in a supervised way to instruct the algorithm to find specific patterns. In pharmaceutical research, these algorithms can identify compounds that have similar properties to those that have worked in the past, helping to select candidates. In addition, AI can be used to create digital twins of experiments, predicting outcomes and helping with study design. AI offers numerous opportunities to improve processes in pharmaceutical research.

Vendela: Thanks, Thomas. Stephan, could you share your insights on centralizing datasets for analytics and how it helps your processes?

Stephan: Centralization of data sets is crucial for efficient data management. In pharmaceuticals, patents are based on chemical structures; these and their associated libraries are of significant value and are typically stored and protected internally. However, AI can clearly help identify hits and reject non-viable compounds in these internal datasets. With publicly available databases, there are limitations due to the high cost of accessing large data sets. Retrieving and downloading data from these databases often requires significant financial investment. Nevertheless, AI can help search for patents and identify trends from various data sources, which can contribute to target identification and analysis.

Vendela: Thank you, Stephan. Thomas, what role do data layers and platforms play in efficient data management, collaboration and decision making in pharmaceutical research?

Thomas: Data layers and platforms are critical to the efficient management and use of data. They help ensure that data is discoverable, accessible, interoperable, and reusable (FAIR principles). By centralizing data and creating semantic interoperability, data platforms enable efficient data integration and analysis. They facilitate collaboration by providing a unified interface for accessing and sharing data across different departments and organizations. With the right data layers and platforms, decision making becomes more informed as data can be abstracted and connected, unlocking new AI use cases and driving innovation.

Vendela: Finally, looking to the future, how do you see AI playing a role in the pharmaceutical industry, particularly in identifying unmet needs for drug development?

Stephan: The future of AI in the pharmaceutical industry is very promising. Currently, AI is used in different stages of drug development, but there is still room for further integration and networking between these stages. For example, AI could play a significant role in patent search, overcoming language barriers and enriching search results. In addition, AI has the potential to leverage diverse data sources, including omics and social media, to identify emerging trends and targets. The challenge is to leverage the immense knowledge and contextual information to enable more holistic decision making and automation.

Thomas: I agree with Stephan. The ongoing development of AI in the pharmaceutical industry is an extraordinarily exciting and transformative time. As AI algorithms continue to improve, their ability to leverage big data is becoming more advanced. This opens up opportunities for personalized medicine, improved decision making and automation. Ongoing efforts to combine multiple data sources and leverage multimodal data are increasing the capabilities of AI systems. The focus on filling knowledge gaps and creating more holistic decision models is paving the way for innovative solutions in drug development.

Vendela: Thank you for your time and valuable insights into the evolving role of AI in the pharmaceutical industry and identifying unmet needs in drug development.