Our main research theme is quantifying and mining the rich information present in cellular images to yield biological discoveries, often using deep learning. We work on high-throughput projects (100,000-1,000,000 images) probing a variety of biological processes and diseases of interest, with a special interest in psychiatric research, infectious disease, and cancer.
High-throughput imaging experiments generate extremely large, multidimensional data sets with quantifiable phenotypic information for every individual cell. Using machine learning, including deep learning, we mine this rich, latent information to identify patterns resulting from chemical or genetic perturbations to probe the causes and cures for various diseases. For example:
- Predicting how new chemical compounds act in cells
- Identifying and classifying toxicity of compounds destined for clinical trials
- Identifying differences in cell structure between patient cells affected by bipolar disorder or schizophrenia
- Discerning the functional impact of gene variants associated with human disease
- Identifying gene function from large-scale genome sequencing studies
We developed the Cell Painting assay in order to carry out high-throughput morphological profiling experiments.
We led the JUMP-Cell Painting Consortium to create the world's largest public Cell Painting dataset of chemical and genetic perturbations.
Impact on human health
Our research has yielded discoveries in several translational projects, some of which have already had a direct impact on the treatment of disease. For example, CellProfiler has been used to identify several small molecules that are effective in treating particular diseases in mouse models. In some cases, discoveries made using CellProfiler have even led to clinical trials in humans, and directly improved patient outcomes [more details]. Our lab's freely available image-based profiling strategies and assays have been adopted by several startups and pharmaceutical companies, one of which (Recursion), now has several drugs in clinical trials.
Community Organizing / Open Source
We helped create academic societies to bring the community together (SBI2)(CytoData). We brought the bioimage software community together in a single online forum (forum.image.sc). We organize public resources of data (BBBC and the Cell Painting Gallery). We organize public data challenges (Data Science Bowl). We launched and lead the NIH Center for Open Bioimage Analysis (COBA) to serve the cell biology community’s growing need for sophisticated software for light microscopy image analysis.
Past Research Areas:
CellProfiler and other bioimage analysis software
We launched and led the open-source CellProfiler project for 18 years. It is beloved by tens of thousands of biologists around the world. Since 2021, the Cimini lab is leading development (CellProfiler site).
CellProfiler Analyst: Machine learning for high-content screens
We launched and led the open-source CellProfiler Analyst project for 16 years. It helps biologists to train supervised machine learning algorithms to identify complex phenotype in high-throughput microscopy experiments. Since 2021, the Cimini lab is leading development (CellProfiler Analyst site).
Quantifying dynamic phenotypes
Many biological questions can only be investigated by collecting time-lapse movies. We have analyzed these images to identify, for example, novel cell cycle landmarks and motor protein regulators. We have also integrated this data with flow cytometry data to quantify unusual cell cycle outcomes.
Imaging flow cytometry
Imaging flow cytometry combines the high-throughput nature of flow cytometry with the high-resolution nature of fluorescence microscopy. For each experimental sample, it yields hundreds of thousands of images of individual cells. We are developing methods to mine these large datasets [NSF project page].
In co-cultured cell systems, two or more cell types are grown together in order to maintain more native physiological functions, enabling experiments that test genetic and chemical perturbations in a more realistic environment. We have developed image analysis approaches to extract information from fluorescence microscopy images of these cell systems, enabling experiments in liver regeneration and hepatotoxicity [NSF CAREER project page].
Quantifying C. elegans
The worm C. elegans can be robotically prepared and imaged and is an effective model to probe a variety of biological questions that require whole animals rather than isolated cells. We have developed sorely needed C. elegans analysis algorithms and validated them in specific large-scale experiments to identify regulators of fat metabolism and pathogen infection.