Our laboratory wants to understand how cells divide and proliferate. We want to discover new biochemical mechanisms that help cells divide when needed. In the same manner, we want to learn how cells stop whenever cell division is too dangerous and could result in irreversible cellular damage. In humans, problems in the control of cell division cause diseases such as cancer and can interfere with wound healing or the maintenance of adult stem cells. In addition, for many bacteria, parasitic organisms, and agricultural pests, the capacity to stop cell division and enter dormant or quiescent states is critical for becoming resistant to antibiotics and pesticides. Thus, understanding how cells activate or stop their cell division machinery promises to advance both biomedical and agricultural knowledge.

To analyze dividing and dormant cells, we use a unique combination of computer vision, genetics, and biochemistry. We track individual cells as they enter or exit from cell division using custom-made deep learning pipelines for image analysis. The information derived from monitoring single cells is processed using machine learning algorithms to cluster data sets, identify correlations, and infer causality in intracellular biological networks. Machine learning-inspired hypotheses are then tested using biochemical and genetic tools in model fungal organisms such as Saccharomyces cerevisiae, Colletotrichum acutatum, Verticillium dahliae, Ustilago maydis, and more.

• Understanding how cells stop their cell cycle machinery and enter an intriguing state called “cellular quiescence or dormancy”

• Implementing deep-learning and generative AI approaches to identify, track, and predict the fate of single cells

• Developing new genetically encoded sensors to track intracellular processes during entry and exit from cell division

• Engineering experimental setups for observing entire signaling pathways, protein networks, and life cycles in single cells

Figure 1. We teach computers to recognize microbes on images and videos to better study them, for instance in this video of Colletotrichum acutatum, a plant parasite, dormant spores germinate and produce a complex labyrinthic form called mycelium. Right: time lapse microscopy video. Center: algorithmic deep learning detection of fungal cells. Left: identified cells with pink contour.

Figure 2. This video shows growing microcolonies of the single celled-fungus, Saccharomyces cerevisiae, which is used to produce bread and wine.  The green masks on top of some cells represent how a computer algorithm learns to detect yeast cells that are mating. 

Figure 3.  In this video, a cell mask-overlap algorithm tracks a single yeast cell, the one with the green halo, as it duplicates itself to produce a clonal microcolony (descended from one single cell).