Dry cell biology

The increasing complexity and volume of microscopy data presents an opportunity for deciphering fundamental biological processes by analyzing image data. Discovery of complex dynamic patterns in such data require deep knowledge of the biological process, the experimental possibilities, the type of information that can be extracted and the computational tools to extract it. The field is desperately in need of computational (or ‘dry’) cell biologists, interdisciplinary scientists who can drive biological inquiry by bridging cell biology and computer science.

We work at the boundaries of quantitative disciplines and cell biology, identify scientific problems, steer (but usually do not perform) experiments, import, adapt and apply quantitative tools, and interpret the data to conceive testable numerical predictions. Read our "dry cell biologist" essay here and our perspective on data science in cell imaging here.

 

Visual interpretability of deep learning models

 

Deep learning has emerged as a powerful technique to identify hidden patterns in complex cell imaging data, but is criticized for lacking the ability to offer meaningful explanations of which image properties drive the network prediction. We are developing methods to "reversed engineer" neural networks by generating synthetic images in a controlled manner to amplify subtle image properties that critically define the model’s classification (Viewpoint) and demonstrate applications in predicting melanoma metastatic aggressiveness (paper, The Scientist highlight, twitter summary), in vitro fertilization, and image-image translation.

LCH_zoom_v3_long.gif
 

Bottom-up modular characterization of tissue state

How multicellular patterns emerge from the heterogeneous behavior and interactions of individual cells? We are developing data science methodologies for deciphering the information propagation between single cells in multicellular systems, driving an emergent collective cell intelligence, and applying it to diverse biological systems such as long-range cell-cell mechanical communication (biorxiv, twitter summary), collective cell death (paper, twitter summary), multicellular calcium synchronization (paper, twitter summary), and tissue organization.

BBsOfCollectiveCellBehavior.png

Machine learning applied to in-vitro fertilization (IVF)

Automated live embryo imaging has transformed IVF into a data-intensive field leading to the development of unbiased and automated methods that rely on machine learning for embryo assessment. Our contributions are in decoupling AI-based implantation prediction and embryo ranking (medrxiv, twitter summary), revealing that phenotypic correlations between sibling embryos provide predictive value regarding an embryo’s future implantation potential (biorxiv, twitter summary), visually interpreting the subtle encapsulated embryonic patterns identified by deep neural networks.

2_developingEmbryo.gif
 

Promoting computational cell biology by sharing and reusing cell image data

A key step toward involvement of computational scientists in cell biology is making image cell data publicly available for secondary analysis, and appreciation of such research by the cell biology community. As the first massive datasets are becoming publicly available we are now at the dawn of this revolution (Read our Perspective).

We are applying "new" data science approaches on "old" live cell imaging data to learn new biology. Specific project include protein localization and high content phenotypic screening. 

bridging_MBoC-01.jpg