Anne Carpenter - Abstract and Bio
Abstract
Gene module discovery by cytological profiling of RNAi-perturbed cells
Anne E. Carpenter
Broad Institute, Cambridge, MA 02142.
Many well-studied biological pathways were originally identified by observation of a mutant phenotype (that is, an observable trait) in a model organism. Even if the biological significance of the phenotype was unknown, identification of the genes related to the phenotype formed the basis of a genetic module for further study. Identifying genes linked to an unusual or uncharacterized phenotype is a relatively unbiased way to open up new biological fields of study and characterize the functions of previously uncharacterized genes in the genome.
We therefore rapidly identified multiple gene modules in human cells based on phenotypes induced by RNA interference (RNAi, a method of genetic perturbation) in a single experiment that was originally designed to identify mitotic regulators. In this approach, biologists train a computer to recognize cellular phenotypes of interest (often quite subtle) in fluorescence microscopy images. Machine learning algorithms then distinguish cells of interest based on each cell’s cytological profile, its rich set of image cytometry-measured features including size, shape, intensity and texture. Genes within the same genetic module often produce cells with a particular rare phenotype (often less than 1 in a 100,000 cells in a large screen).
We rapidly scored millions of individual cells for ~15 phenotypes, yielding multiple predicted gene modules that we are now validating and pursuing. The methods developed for this project are incorporated into the soon-to-be-released, open-source software, CellVisualizer (www.cellvisualizer.org), which allows biologists to conveniently score high-throughput screens for complex and subtle visual phenotypes.
Bio: Anne Carpenter, Ph.D.
Dr. Carpenter is the Director of the Imaging Platform at the Broad Institute of Harvard and MIT, in Cambridge, MA. Trained as a cell biologist, she now leads an interdisciplinary research group that develops and applies methods for extracting quantitative information from biological images. The resulting software, CellProfiler and CellVisualizer, is released open-source to the scientific community.
Anne earned her B.S. from Purdue University in 1997 and her Ph.D. from the University of Illinois at Urbana-Champaign in 2003. During her postdoctoral work with David Sabatini at the Whitehead Institute for Biomedical Research, she was co-mentored by Polina Golland at the Computer Science/Artificial Intelligence Laboratory at the Massachusetts Institute of Technology.
She has been awarded fellowships and awards from the Howard Hughes Medical Institute, National Science Foundation, Computational and Systems Biology initiative, the Life Sciences Research Foundation, and the Society for Biomolecular Screening.
