Postdoctoral researcher Iguaracy Pinheiro de Sousa is using bioinformatics to support the development of better organoids. Working at EBI in the Petsalaki research group, Pinheiro de Sousa uses single-cell transcriptomics data to understand how different types of cells interact and communicate to form a tissue. One of the applications of his work is the development of organoids – miniaturised 3D tissue cultures, derived from stem cells, and designed to mimic the structure and function of organs.
Can you tell us more about your background in biomedical science and how it led to your current research focus?
Pinheiro de Sousa: I chose Biomedical Science mainly for research purposes. When choosing my Bachelor’s course after high school, I was wavering between Medical School, Biomedicine, and Biology. Medical School was too clinic-focused, while Biology was too broad. I found Biomedicine to be a middle ground and became specifically interested in human genetics.
I pursued a Master’s in Genetics and Molecular Biology, working on a small study focused on pharmacogenetics in a hypertensive cohort from the Amazonian population, exploring how their genetic background influenced a particular hypertensive drug treatment. This led me to continue working on cardiovascular disease and eventually move to the Heart Institute in São Paulo.
What inspired your transition from wet lab experiments to computational bioinformatics work?
Pinheiro de Sousa: I started encountering questions that required a more holistic approach, which was often limited by wet lab experiments and time-consuming methods. Bioinformatics allowed me to address these questions faster and with a broader perspective. The growing availability of public data facilitated this. However, the transition wasn’t easy. I tried many online courses, but the turning point was having a bioinformatics project. Through their Next Generation Scientist program, I was selected for an internship at Novartis, in Basel, Switzerland. For three months away from the bench, I focused on developing computational skills in RNA-seq and ChIP-seq data analysis. Engaging in a bioinformatics project was key to enabling my transition.
My PhD research on cardiovascular disease was incredibly helpful, especially in interpreting computational results and extracting insights from big data. This knowledge helps translate findings into strategies for building better cardiac organoids; 3D miniaturised, cell-based in vitro models designed to mimic the structure and function of organs.
Can you explain how you are using bioinformatics and single-cell transcriptomics data to improve the development of organoids?
Pinheiro de Sousa: I am using single-cell transcriptomics data to predict interactions between different cell types in heart tissue. By integrating databases based on known protein-protein interactions and prior knowledge of which proteins act as receptors and ligands, we can predict how cells communicate based on the gene expression levels of a receptor in one cell type and a ligand in another. This approach provides insights into higher-order cell-cell communication within a tissue, helping to identify the most relevant cells to co-culture with cardiac cells, thereby enhancing organoid complexity. Initiatives like the Human Cell Atlas (HCA) project, which aims to map all human cell types, provided the heart cell atlas data that made this work possible. Bioinformatics plays a key role by leveraging this data for biomedical applications.
Current organoid models often lack cellular diversity, spatial organisation, and maturity to replicate native tissues. They tend to replicate only certain aspects of tissues, failing to capture the full complexity of cell types, maturation levels, and functions. Better organoids could address these gaps by providing models that more accurately mimic the native tissue environment, enabling a deeper understanding of tissue complexity and functionality.
My project aims to address these limitations by developing a data-driven strategy to guide organoid construction. By selecting the optimal cell type combinations based on their interactions in native environments, we could improve cellular diversity and maturity. The ultimate goal is to create a user-friendly computational tool for generating data-driven organoid construction hypotheses, using cardiac organoids as a proof of concept.
How do you envision better organoids contributing to drug discovery and medical research?
Pinheiro de Sousa: Breakthroughs in organoid technology could lead to more accurate and functional models, supporting safer and more effective clinical research. Regulatory shifts, such as the FDA allowing clinical trials without prior animal testing, highlight the urgency for such advancements. Improved cardiac organoids, for example, could be used in ischemic heart disease, which leads to the loss of billions of cardiomyocytes and has limited regenerative treatment options. Thus, organoids that are more like adult tissue could serve as physiologically relevant platforms for drug testing, disease modelling, and regenerative medicine, key for medical research and therapeutic development.
How does your work on nanorobots tie into your research on organoids and precision medicine?
Pinheiro de Sousa: My work primarily focuses on finding ‘addresses’ in the vasculature rather than developing the nanorobots themselves. The vasculature forms a continuous pathway that reaches all parts of the body. Every living cell is located within a small radius (100 µm) of the nearest blood vessel to receive oxygen, with few exceptions. A proof of concept using the vasculature for delivery was demonstrated by Yuliang Zhao et al. (2018), where DNA origami nanorobots were designed to deliver thrombin, a blood coagulation protease, to tumour-associated endothelial cells via the surface protein nucleolin. Nucleolin functioned as both a targeting domain and a trigger for the nanorobot thrombin release.
This approach successfully induced intravascular thrombosis, causing tumour necrosis and inhibiting tumour growth in mice. Connecting this to my research on organoids, if precise drug delivery is to be tested in humans in the future, we will need more robust and physiologically relevant organoids, particularly vascularised ones, to serve as platforms for testing these delivery systems.
How does the collaboration between computational and wet lab scientists accelerate organoid research?
Pinheiro de Sousa: When I moved to EMBL-EBI, I used to think computational scientists were at a disadvantage because they were less involved in experimental design (although this has changed significantly in recent years), while wet lab scientists were the main drivers of experimental design and data generation. However, computational work provides valuable insights that can generate new hypotheses to guide data generation, making it essential for these two fields to collaborate more closely. Without these wet-dry lab collaborations, we miss opportunities and place science as a whole at a disadvantage. Greater collaboration, particularly on hypotheses that require expertise in both fields, would unlock significant benefits for research. In the context of organoids, this collaboration is especially critical, as it integrates computational models with experimental data. This partnership enables the rapid testing of hypotheses and improves the design of organoids by combining computational predictions with biological experiments.
How will bioinformatics and data science advancements impact research in the coming years?
Pinheiro de Sousa: Advancements in bioinformatics and data science are already profoundly influencing biological research, but their potential impact could be even greater if the communication gap between computational and experimental scientists is bridged. These two groups often "speak different languages," making active and collaborative communication essential for progress.
Better dialogue is key to formulating relevant questions from the massive amounts of data generated daily; questions that are computationally insightful and experimentally testable in the wet lab. However, some biological problems are difficult or impractical to explore experimentally. With the right data, such challenges can be addressed through computational modelling, enabling in silico simulations of complex biological processes. This approach saves time and resources while opening avenues to explore otherwise inaccessible scenarios.
The synergy between computational and experimental biology is crucial for increasing impact in biological research. Computational models can guide wet lab experiments, while experimental results refine and validate computational predictions. This back-and-forth collaboration creates a feedback loop that can drive transformative discoveries and advance our understanding of complex biological systems.
Can you describe some of the collaborative projects you are involved in, both within EMBL?
Pinheiro de Sousa: Some questions in science require a highly interdisciplinary approach, and it can be challenging for one person to understand all the areas needed to address such questions deeply. This is where collaborations become crucial. Collaborating with experts from different fields broadens your perspective and guides your project in new directions while still addressing the core question. I continue collaborating with my former supervisor from Brazil, where all the cardiac organoid experiments are being tested. My co-supervisor is based at EMBL-Barcelona, developing vascularised tissues in vitro.
Recently, we started a collaboration with a nanoparticle engineer from Manchester University, which will be relevant for downstream projects involving targeted delivery systems. One particularly enjoyable collaboration was with a postdoc I met at the Network Signaling conference in Cold Spring Harbor, NY. She is based at Yale University and was investigating interactions between neurons and oligodendrocytes in healthy and Alzheimer’s brains. Using a combination of cell-cell communication analysis pipeline and subcellular proteomics, we identified distinct communication patterns in Alzheimer’s brains. This fruitful collaboration generated two works that have just been accepted, one in Nature Neuroscience and the other in Nature Ageing. This partnership was unique because I contributed to answering her questions rather than the other way around, which is how collaborations should ideally work—mutually beneficial in both directions.
What advice would you give to young scientists interested in interdisciplinary research like yours?
Pinheiro de Sousa: My advice would be to embrace curiosity and actively seek collaborations across different fields, as these experiences can broaden your perspective. Interdisciplinary research thrives on combining diverse perspectives, so don’t hesitate to engage with experts outside your immediate area of expertise. Attend talks and lectures beyond your field. They can inspire unexpected connections, as they did for me during my PhD. Building a solid foundation in your primary discipline is essential, but developing communication skills to bridge gaps between fields is equally important.
Finally, stay open to continuous learning and adaptation, even though it can be challenging. It’s difficult to fully grasp other fields' methods and limitations, which can be frustrating. While you can’t master everything, you can learn enough to address your research questions meaningfully. This is why collaboration is crucial. Experts in other fields can guide you on what’s possible and help focus your efforts, making interdisciplinary work productive and transformative.
Iguaracy Pinheiro-de-Sousa is a Postdoctoral Fellow at EMBL’s European Bioinformatics Institute (EMBL-EBI).
