09-11, 09:45–10:00 (Europe/Berlin), Main conference room - MW 0350
Reconstructing temporal cellular dynamics from static single-cell transcriptomics remains a major challenge. Methods based on RNA velocity are useful, but interpreting their results to learn new biology remains difficult. Here we show NeuroVelo, a method that couples learning of an optimal linear projection with non-linear Neural Ordinary Differential Equations. Unlike current methods, it uses dynamical systems theory to model biological processes over time, hence NeuroVelo can identify what genes and mechanisms drive the temporal cellular dynamics. We benchmark NeuroVelo against several state-of-the-art methods using single-cell datasets, demonstrating that NeuroVelo has high predictive power but is superior to competing methods in identifying the mechanisms that drive cellular dynamics over time. We also show how we can use this method to infer gene regulatory networks that drive cell fate directly from the data.
Understanding temporal cellular dynamics from static single-cell transcriptomics data remains a significant challenge in computational biology. Traditional RNA velocity-based methods, while useful, often fall short in interpretability. In this study, we introduce NeuroVelo, a novel approach that integrates the learning of optimal linear projections with the flexibility of non-linear Neural Ordinary Differential Equations (Neural ODEs). By leveraging dynamical systems theory, NeuroVelo models biological processes across time, enabling the identification of key genes and mechanisms driving cellular dynamics.
NeuroVelo stands out with its high predictive accuracy and superior capability in uncovering the underlying mechanisms of cellular behavior compared to existing methods. Furthermore, our method facilitates the direct inference of gene regulatory networks from single-cell data, offering deeper insights into cell dynamics. This presentation will go through the methodology, demonstrate its predictive aspect as the other methods, and showcase its applications in uncovering new biological insights.
Previous knowledge expected
Idris graduated from the University of Science and Technology Houari Boumediene in Algiers with a master degree in theoretical physics, then obtained a diploma in quantitative life sciences from the Abdus Salam International Centre for Theoretical Physics (ICTP).
He is currently working on using neural ordinary differential equations for systems biology with Guido Sanguinetti and Andrea Sottoriva.