Sponsored Content by Axxam S.p.AReviewed by Olivia FrostJul 23 2026
In this interview, News Medical speaks with Stephan Brock, Alessandro Cinti, and Samuel Bradberry about how AI-powered analytics and experimental validation are transforming target identification in drug discovery. The discussion highlights how Molecular Health and Axxam combine computational analysis and experimental validation to support the identification and prioritization of promising therapeutic targets. They discuss the importance of transparent, evidence-driven decision-making, multimodal data integration, disease-focused target prioritization, and the ways in which combining computational insights with biological validation can improve confidence in therapeutic development and downstream drug discovery success.
Can you please introduce your roles at Molecular Health and Axxam?
Stephan Brock: Our focus is on applying AI and advanced analytics across the drug development lifecycle, from target identification and indication finding through to safety analytics, clinical trial optimization, and safety prediction. My work is centered on building evidence-based analytical frameworks that help transform biological data into actionable decisions for drug discovery.
Alessandro Cinti: My role involves supporting the transition from computational target identification into experimental validation. We work across a broad range of therapeutic areas, developing disease-relevant biological models and validation strategies that help establish the biological relevance and therapeutic potential of candidate targets.
Samuel Bradberry: My work focuses on integrating computational and experimental approaches across the early stages of drug discovery. A key area of interest is ensuring that the data generated during target identification can be translated into meaningful biological validation and ultimately support successful drug discovery programs.
How is AI-powered analytics changing target identification in drug discovery?
Stephan Brock: AI is now being applied throughout the drug development process, and in several areas, it is already demonstrating tangible value. Molecular design and pattern recognition are areas where AI has shown significant promise, and we are now seeing AI-supported drug candidates advance to advanced stages of clinical development.
However, despite these advances, attrition rates across drug development remain high. Critical decisions still need to be made around mechanisms of action, patient populations, efficacy, safety, and risk-benefit assessments. AI can help generate signals and identify patterns, but the real challenge lies in translating those signals into decision-grade evidence that can support drug development.
Why is drug discovery fundamentally a decision-making challenge rather than simply a prediction problem?
Stephan Brock: Drug development operates under considerable uncertainty. While predictive models are valuable, the ultimate goal is not simply to predict outcomes but to make informed decisions based on robust evidence.
Many AI systems excel at understanding context and generating signals, but they often lack the transparency and traceability required to support high-stakes scientific decisions. At Molecular Health, we focus on ensuring that every insight is grounded in evidence that can be reviewed, challenged, and validated. The objective is to move beyond prediction and support better decision-making throughout the development process.
How does Molecular Health use multimodal data and knowledge graphs to support target identification?
Stephan Brock: One of the biggest challenges in drug development is the fragmented nature of biomedical data. Information exists across clinical datasets, experimental studies, omics data, literature, diseases, and drugs. If these sources remain disconnected, it becomes difficult to generate actionable insights.
To address this, we have built a deeply integrated multimodal data environment implemented through a knowledge graph. This allows us to preserve biological context and represent how different entities relate to one another. Rather than simply aggregating information, we create a connected framework that reflects the complexity of underlying biology and enables more meaningful interpretation of disease mechanisms and therapeutic opportunities.
How do you move from target prediction to target prioritization?
Stephan Brock: Our objective is not merely to generate hypotheses but to prioritize targets that can support informed decision-making. We do this by grounding each candidate in multiple independent evidence streams, including genetics, gene expression, clinical observations, disease associations, literature evidence, and knowledge graph-derived relationships.
Importantly, these evidence layers are intentionally orthogonal, meaning they provide independent perspectives on the same biological question. Strong targets emerge when multiple lines of evidence converge consistently. We then evaluate factors such as biological context, mechanistic relevance, tractability, and validation readiness. The result is a ranked set of target candidates supported by transparent and traceable evidence.
How are promising targets prioritized and transitioned into experimental validation?
Alessandro Cinti: The data generated during target identification plays a critical role beyond ranking targets. We use this data to contextualize candidates within disease-relevant pathways, identify potential translational readouts, assess tractability, and anticipate possible safety liabilities.
Moving from a longlist of candidates to a refined shortlist is a collaborative, data-informed process carried out with Molecular Health. Once a final target has been selected, we begin experimental validation with the aim of establishing both biological relevance and mechanistic involvement within disease-relevant systems.
What are the key pillars of Axxam’s target validation strategy?
Alessandro Cinti: Our validation strategy is built around three complementary pillars that allow us to interrogate target biology from multiple perspectives. First, we determine whether the target is expressed in relevant biological systems, including established cell lines, primary cells, induced pluripotent stem cell-derived models, and, when available, patient-derived samples.
Second, we apply genetic perturbation approaches, such as CRISPR-Cas9, CRISPR interference and activation, and RNA interference, to assess causality between target modulation and cellular phenotype. Third, we incorporate phenotypic profiling and functional assays to understand the downstream biological consequences of target modulation and to build a comprehensive mechanistic understanding of the target.
How do genetic perturbation, phenotypic profiling, and functional assays strengthen target validation?
Alessandro Cinti: These approaches allow us to move beyond simple observations and establish causal relationships between a target and disease-relevant phenotypes.
We employ multiparametric phenotypic profiling technologies such as high-content imaging and cell painting to capture complex biological responses, including morphological changes and organelle-level alterations. Depending on the therapeutic area, we may also integrate electrophysiological measurements, transcriptomics, proteomics, and custom functional assays that investigate enzyme activity, receptor signaling, protein-protein interactions, or other disease-relevant mechanisms.
By combining these diverse datasets, we can generate a coherent, multidimensional understanding of target biology while increasing confidence in its therapeutic potential.
What disease-relevant models and assay platforms can be used to validate targets across different therapeutic areas?
Alessandro Cinti: The appropriate biological context always depends on the indication being studied. Over the years, we have developed expertise across a wide range of disease models and assay platforms.
For cardiovascular research, we integrate fluorescence imaging, calcium oscillation measurements, and electrophysiological systems to evaluate functional outcomes. In neuroscience, we combine calcium imaging, voltage indicators, patch-clamp methodologies, and induced pluripotent stem cell-derived neuronal systems to study neuronal activity and network behavior.
In oncology settings, we use recombinant tumor cell lines, patient-derived samples, and increasingly sophisticated three-dimensional tumor models. We also work with retinal cells, fibroblasts, keratinocytes, melanocytes, intestinal models, and many other disease-relevant systems. These platforms allow us to tailor validation strategies to the specific biology of each target and therapeutic hypothesis.
Why is the integration of AI-powered analytics and experimental validation so important for future drug discovery programs?
Samuel Bradberry: The greatest value emerges when target identification and target validation are approached as a continuous, integrated process rather than as isolated activities.
Computational prediction alone is not sufficient. To transform predictions into meaningful outcomes, they must be experimentally validated in relevant biological systems. Likewise, experimental data can be used to refine computational models and generate new evidence. This creates a continuous feedback loop that strengthens confidence in decisions throughout the discovery process.
These early stages serve as the foundation for everything that follows, from hit identification and lead optimization through to broader drug discovery programs. By combining AI-powered analytics with rigorous validation, we can generate higher-quality evidence, improve decision-making, and ultimately increase the likelihood of successful therapeutic development.
About Stephan Brock
Stephan Brock, PhD, is co-founder and CTO of Molecular Health. He leads the development of AI-driven, knowledge graph–based systems for decision support in pharmaceutical R&D, integrating molecular, clinical, and real-world data into actionable insights. His work spans bioinformatics and systems biology, including regulatory-grade applications under the IVDR, data-driven solutions for the FDA, and the design and implementation of clinical data infrastructure at leading cancer centers. He holds a PhD in Biochemistry from the University of Bayreuth.
About Alessandro Cinti
Alessandro Cinti is a Scientific Research Expert at Axxam with over 15 years of experience spanning international academic research and industrial biotech R&D. He currently specializes in drug discovery and early-stage development, with a focus on small molecules, innovative platforms, and translational science. Prior to joining Axxam, Alessandro conducted postdoctoral research at McGill University, focusing on the molecular biology of HIV-1 and host–pathogen interactions in Ebola infection. He holds a PhD in Translational and Molecular Medicine and is the author and co-author of several scientific publications in the field.
About Samuel Bradberry
Samuel (Sam) Bradberry is a Principal Scientist in the Data Sciences team at Axxam. He supports data and knowledge-driven discovery for Axxam’s clients and the development and management of internal data and knowledge infrastructures. He holds a PhD (2017, Trinity College Dublin) and a Master’s degree in chemistry (2012, University of Southampton), and has specialized throughout the last decade in management, processing, and interpretation of scientific data in the chemical and life sciences. Sam's current interests are in the close integration of data, knowledge, and in silico methods with wet-lab experimental cycles to enrich the discovery process.
About Axxam S.p.A
Axxam S.p.A. is a Contract Research Organization (CRO) providing integrated early discovery services to the Life Sciences sector. We support pharmaceutical and biotechnology companies, start-ups, patient foundations, and academic institutions throughout their journey from target identification to lead generation, across all therapeutic areas and target classes.
Axxam is a biology powerhouse, translating innovative science into bioactive molecules. Our integrated target-based drug discovery offering, known as DiscoveryMAXX, spans target identification and validation, assay development, hit identification through high-throughput screening using Axxam’s high-quality synthetic and natural compound collections or client-provided libraries, hit validation, hit-to-lead activities, and compound management.
DiscoveryMAXX is founded on the principle that better decisions early in drug discovery significantly reduce the risk of downstream failure, delays, and budget overruns during preclinical and clinical development.
In addition to pharmaceutical and biotechnology projects, we apply the same rigorous scientific approach to identify novel bioactive compounds for crop protection, animal health, food and beverage, pet food, cosmetics, and fragrance industries.
For more information, please visit www.axxam.com.
Capabilities and expertise
Axxam's capabilities cover the full continuum of early discovery, including:
- Molecular biology (cloning, gene expression analysis, protein expression, genomic engineering)
- AI-enhanced target identification and wet-lab validation
- Assay development (cellular and biochemical)
- Compound management
- Fully automated high-throughput screening, with diverse readouts including electrophysiology, quantitative gene expression, and high-content imaging
- Virtual screening
- Hit characterization and profiling
- Access to Axxam’s physical compound collections (universal and focused) and virtual libraries
- Hit-to-lead optimization
- Early in vitro assessment of efficacy, biodistribution, and toxicity, including New Approach Methodologies (NAMs)