Uploaded May 2026 | Updated September 2026, 2 weeks ago
Eric and Wendy Schmidt Center Symposium: Biomedical Science and AI
April 28 - 29, 2026
Day 2, Invited talk: Interpretable Representation Learning for Wound Healing Dynamics
Bianca Dumitrascu, Columbia University
Abstract:
Single-cell RNA-seq enables the study of cell states across diverse biological conditions, such as aging, drug treatments, and tissue injury. However, disentangling shared and condition-specific transcriptomic patterns remains a significant computational challenge, particularly in settings with missing data or complex experimental designs. In this talk, I will introduce new deep generative frameworks designed to disentangle these transcriptomic signals, allowing for robust integration, cross-condition prediction, and biologically interpretable insights. Using real and simulated scRNA-seq datasets, we demonstrate that disentanglement can uncover shared wound healing patterns and distinct changes in cell behavior, including age-dependent immune responses and drug-modulated extracellular matrix remodeling. Finally, I will discuss open problems towards a pipeline for synthetic self-organizing systems.
Speaker bio:
Bianca Dumitrascu joined Columbia University in the Spring 2023 as an Assistant Professor of Statistics and Herbert and Florence Irving Assistant Professor of Cancer Data Research (in the Herbert and Florence Institute for Cancer Dynamics and in the Herbert Irving Comprehensive Cancer Center). Before Columbia, Bianca was an Affiliated Lecturer in the Department of Computer Science and Technology (Computer Laboratory) at the University of Cambridge. Bianca was a Departmental Early Career Fellow in the Accelerate Programme for Scientific Discovery. Bianca works at the intersection of machine learning and genetics. Her main research interest is understanding how local molecular rules give raise to emergent spatial patterns in the context of biological dynamical systems. To this end, Bianca uses techniques from statistical optimization, statistical physics and domain adaptation to identify contextual phenotypes in spatial transcriptomic data and to understand the identity of single cells and their interactions in early developement. She is also interested in active learning and graphical neural networks as models to study message passing in multi-agent systems.
Previously, Bianca was a Member in the School of Mathematics at the Institute for Advanced Study and she attended the semester long program in deep learning at the Statistical and Applied Mathematical Sciences Institute, where she was graciously hosted by David Dunson in the Duke Statistical Science Department. She received my PhD in Computational Biology at Princeton University, under the mentorship of Barbara E. Engelhardt. Her PhD research focused on the effect of experimental design in single cell gene expression studies and on method development for structured, high-dimensional medical and genomic data. Bianca did her undergraduate studies in Mathematics at MIT.
For more information, visit broad.io/ewsc-symposium-2026
Copyright Broad Institute, 2026. All rights reserved.
Eric and Wendy Schmidt Center Symposium: Biomedical Science and AI
April 28 - 29, 2026
Day 2, Invited talk: Interpretable Representation Learning for Wound Healing Dynamics
Bianca Dumitrascu, Columbia University
Abstract:
Single-cell RNA-seq enables the study of cell states across diverse biological conditions, such as aging, drug treatments, and tissue injury. However, disentangling shared and condition-specific transcriptomic patterns remains a significant computational challenge, particularly in settings with missing data or complex experimental designs. In this talk, I will introduce new deep generative frameworks designed to disentangle these transcriptomic signals, allowing for robust integration, cross-condition prediction, and biologically interpretable insights. Using real and simulated scRNA-seq datasets, we demonstrate that disentanglement can uncover shared wound healing patterns and distinct changes in cell behavior, including age-dependent immune responses and drug-modulated extracellular matrix remodeling. Finally, I will discuss open problems towards a pipeline for synthetic self-organizing systems.
Speaker bio:
Bianca Dumitrascu joined Columbia University in the Spring 2023 as an Assistant Professor of Statistics and Herbert and Florence Irving Assistant Professor of Cancer Data Research (in the Herbert and Florence Institute for Cancer Dynamics and in the Herbert Irving Comprehensive Cancer Center). Before Columbia, Bianca was an Affiliated Lecturer in the Department of Computer Science and Technology (Computer Laboratory) at the University of Cambridge. Bianca was a Departmental Early Career Fellow in the Accelerate Programme for Scientific Discovery. Bianca works at the intersection of machine learning and genetics. Her main research interest is understanding how local molecular rules give raise to emergent spatial patterns in the context of biological dynamical systems. To this end, Bianca uses techniques from statistical optimization, statistical physics and domain adaptation to identify contextual phenotypes in spatial transcriptomic data and to understand the identity of single cells and their interactions in early developement. She is also interested in active learning and graphical neural networks as models to study message passing in multi-agent systems.
Previously, Bianca was a Member in the School of Mathematics at the Institute for Advanced Study and she attended the semester long program in deep learning at the Statistical and Applied Mathematical Sciences Institute, where she was graciously hosted by David Dunson in the Duke Statistical Science Department. She received my PhD in Computational Biology at Princeton University, under the mentorship of Barbara E. Engelhardt. Her PhD research focused on the effect of experimental design in single cell gene expression studies and on method development for structured, high-dimensional medical and genomic data. Bianca did her undergraduate studies in Mathematics at MIT.
For more information, visit broad.io/ewsc-symposium-2026
Copyright Broad Institute, 2026. All rights reserved.
