Periodic Reporting for period 1 - Cells2Tissues (Cellular models for tissue function in development and ageing)
Berichtszeitraum: 2023-06-01 bis 2025-11-30
Cells are remarkably potent entities that are capable of making complex molecular decisions toward activating specific genes and facilitating specific functions. Such decisions are ultimately derived in ensembles of cells, that together form tissues and organs. In Cells2Tissues we aim at very high resolution profiling and modelling of these fundamental processes, which are responsible to such dramatic events as the emergence of embryos from a fertilized oocyte or essential long-term physiological functions such as those generating our blood and immune cells from stem cells in the bone marrow. To do this, we combine the development of new and rich computational models with experiments that capture single cells and profile their molecular states. In cells2Tissues we have already constructed two main detailed models of the collective functioning of cells – one describing the developing mouse embryo and the other showing how human blood production variate among healthy and ageing individuals. Based on these models we build computational models to understand how cells make decisions – for example by deciphering the code controlling regulation of gene expression, and by inferring how cells are affected by the temporal and spatial position during complex coordinate tissue differentiation and morphogenesis.
Our team in Cells2Tissues is collecting a multi-modal dataset that capture the molecular activity of cells during embryonic and hematopoietic differentiation. These data becomes rapidly richer as technology is improving. We then devise computer algorithms to projects the different data sources on one coherent model. We aim at models that are mechanistic and principled – for example describing explicitly how certain protein binds DNA and interact with each other, or how cells move from the embryo head to its tail. We continuously contrast our bottom-up computational models with models that process massive datasets blinding using a type of AI tools called foundation models. In our project, the ultimate aim is mechanistic understanding and experiment proving it – and the continuous improvement in both bottom up and AI models provide a new powerful research veichle for deriving such understanding.
With the next steps in cells2Tissues, we hope to complete and sharpen our models such that it will be possible to use them for both prediction and design of interventions. In embryonic development, we wish to facilitate highly controlled differentiation of cells toward fates of interest, while imitating as closely as possible the natural differentiation trajectory. We believe that as our models become more and more realistic, this will become feasible. In our analysis of hematopoiesis, we can already diagnose emerging disease based on the aberrations in cell states in circulating stem cells. We will continue to enhance this capability, and in parallel work toward the ability to predict how therapeutic (e.g. hormonal, or those involving targeting of specific molecules) can change the course of emerging disease