Introduction to cell therapies
Recent years have demonstrated the considerable potential of regenerative medicine. Cell therapies, in particular, hold great promise for transforming and improving the lives of patients suffering from various diseases and conditions, including heart disease, diabetes, and neurological disorders.
However, understanding the effect of these therapies requires extensive experimental wet lab assays and the subsequent analysis of generated data. In this context, AI and Machine Learning play a promising role, enabling a deeper and clearer understanding of the panorama to improve wet-lab investment and therapies result.
To improve a therapy, it is necessary to gain a comprehensive understanding of the disease: how it impacts cellular processes and metabolic pathways. This will determine the design of a targeted therapy.
As cell therapies offer a novel and diverse approach for treating multiple conditions and diagnoses across different types of tissues, each one must be treated independently. In this scenario, experimental data availability is limited, and the generation of new experimental data is a costly and time-consuming process. This situation calls for in silico approaches that explore numerous modalities, gaining a clear understanding of the reach and possibilities for the expected treatment before even taking any costly experiments.
Existing biological knowledge is sometimes insufficient to provide the basis needed to develop high-precision cell therapies. In these scenarios, researchers gather information from multiple sources, including papers, reviews, and DNA databases like DDBJ (Japan), GenBank (USA), and European Nucleotide Archive (Europe). These are crucial to designing DNA cell modifications that can result in a novel function of the cell or a redirection of its actions, like CAR-T therapies. While these advances make it easier and cheaper to explore these options, the vast and complex combinatorial space presents challenges that limit the number of wetlab assays that can be tested due to time, resource, and labor constraints.
To effectively navigate this design space, researchers typically vary one component at a time, which can lead to suboptimal results. This could be a cell variant or ambient conditions like pH, temperature, oxygen concentration, cell media, and more. A more efficient approach would be to simultaneously alter multiple variables at the same time, allowing for a broader exploration of potential combinations. This method could help identify optimal treatments more effectively and provide insights into how different components contribute to desired outcomes.
Transomics platforms for cell therapies
A multi-omic comprehension of the cell can come as a solution by linking the in-silico genome, transcriptome, proteome, and metabolome of a cell. This can then be associated to specific cell phenotypes: how the cell looks and works. Therefore, this approach will allow researchers to anticipate the function of yet-to-be-designed cells, allowing an early evaluation before doing tens or hundreds of tests in the laboratory.
To fulfill this goal, we are developing the Multi-omic Network Atlas (MoNA), composed of multi-omic profiles of cell types across ambient conditions structured with system biology knowledge, allowing users to map and understand the relation between the multi-omic phenotype of a cell and its experimental ambient conditions.
This Transomics platform enables the creation of a lab-in-the-loop research system, facilitating an integrated approach to wet lab and in silico research that builds on each other’s findings and enables the optimal sharing of data and insights.
Potential questions to be answered by the tool:
- What are the specific molecular mechanisms underlying the disease? Understanding the pathways and cellular processes that are disrupted in the disease for identifying effective therapeutic targets.
- Which cell types have the potential to effectively modulate these pathways? Determining the most suitable cell types for therapy involves evaluating their ability to differentiate, survive, and perform the desired functions in the context of the disease.
- How will the chosen cells interact with the patient’s immune system and its cellular microenvironment? Understanding the immunological aspects, including potential rejection or immune responses, as well as the interactions with its neighbors, is essential for ensuring the safety and efficacy of the therapy.
- What are the optimal conditions for cell growth, differentiation, and functionality? Identifying the right culture conditions, growth factors, and environmental cues is critical for maximizing the therapeutic potential of the cells.
- How will genetic modifications (if any) affect cell behavior and safety? If the therapy involves gene editing or modification, understanding the implications for cell function, stability, and potential off-target effects is vital for assessing safety and efficacy.
To answer these questions, we are developing different methods that will allow the user to:
Understand the role of each biomolecule in an assay sample for cell analysis and target identification: rapid, large-scale in silico screening of predicted candidates to reduce wet-lab testing.
Create in silico approximations of the cell multi-omic behaviour, bibliography, assays related to your assay of interest, optimal environmental conditions, culture media, or differentiation protocols.
Enhance target discovery. Finding the right targets for gene therapies or drug discovery is complex, as it involves identifying specific genes, proteins, or pathways for effective treatments. Transomics uncovers patterns often missed in manual searches.
Simulate phenotypes: to gain insight into how genetic modifications made to target cells can affect their functionality and the surrounding cellular environment.
Carry out multi-omic data analysis: Transomics supports loading and allows integration of complex data from various omics (genomics, transcriptomics, proteomics, etc.) for holistic analyses.
Screen a large number of candidates rapidly and select designs that fulfill the desired criteria, similar to their use in target identification.
References
Bhandari, M., Chang, A., Devenyns, T., Devereson, A., Loche, A., & Van Der Veken, L. (2022). How AI can accelerate R&D for cell and gene therapies. McKinsey & Company. https://www.mckinsey.com/industries/life-sciences/our-insights/how-ai-can-accelerate-r-and-d-for-cell-and-gene-therapies
Capponi, S., & Daniels, K. G. (2023). Harnessing the power of artificial intelligence to advance cell therapy. Immunological Reviews, 320(1), 147-165. DOI: 10.1111/imr.13236 https://pubmed.ncbi.nlm.nih.gov/37415280/
Nosrati, H., & Nosrati, M. (2023). Artificial Intelligence in Regenerative Medicine: Applications and Implications. Biomimetics, 8(5), 442. DOI: 10.3390/biomimetics8050442 https://pubmed.ncbi.nlm.nih.gov/37754193/
Wu, Y., & Xie, L. (2025). AI-driven multi-omics integration for multi-scale predictive modeling of genotype-environment-phenotype relationships. Computational And Structural Biotechnology Journal. https://doi.org/10.1016/j.csbj.2024.12.030 https://www.sciencedirect.com/science/article/pii/S2001037024004513
Wang, L. L. W., Janes, M. E., Kumbhojkar, N., Kapate, N., Clegg, J. R., Prakash, S., … & Mitragotri, S. (2021). Cell therapies in the clinic. Bioengineering & translational medicine, 6(2), e10214. DOI: 10.1002/btm2.10214 https://pubmed.ncbi.nlm.nih.gov/34027097/
