CSAIL Publications and Digital Archive header
bullet Technical Reports bullet Work Products bullet Research Abstracts bullet Historical Collections bullet

link to publications.csail.mit.edu link to www.csail.mit.edu horizontal line


Research Abstracts - 2006
horizontal line

horizontal line

vertical line
vertical line

Methods for High-Throughput, High-Content Image-Based Screens

Thouis R. Jones (CSAIL), Anne Carpenter (Whitehead Institute for Biomedical Research), InHan Kang (CSAIL), Polina Golland (CSAIL), & David Sabatini (Whitehead Institute for Biomedical Research)


We are developing methods and algorithms for high-throughput image analysis for cell images. These are necessary to enable screens of visual phenotype in genome-scale perturbation experiments. Such methods are also valuable in drug discovery, where large libraries of compounds are screened in a high-throughput manner.

Our image analysis methods are geared towards gathering per-cell data from image-based assays. A single screen can produce hundreds of measurements on millions of individual cells. To explore such large amounts of data, we are working on visualization methods, as well as probabilistic models of cellular variation. Visualization of large cell populations allows a human to detect and separate out interesting outlier subpopulations ("hits") in a screen. Probabilistic methods can automatically find and characterize subpopulations that show a novel phenotype, though further interpretation may require human intervention.

cells segmentation
Drosophila cells stained for Actin,
and their segmentation as found by our algorithm [1].

We are also exploring combinations of automatic and interactive methods, in which a user provides feedback to an online algorithm to build classifiers for cells (novel vs. uninteresting, resting vs. mitotic vs. apoptotic, ...). These classifiers can be used to search for genes or experimental conditions that cause an increase or decrease of a particular phenotype, to guide the user to hits of the desired type. This allows a very targeted exploration of the large amounts of data from image-based screens.

exploring per-cell data

Exploring per-cell measurements from a high-throughput, image-based screen.

[1] Thouis Jones, Anne Carpenter, and Polina Golland. Voronoi-Based Segmentation of Cells on Image Manifolds. In Computer Vision for Biomedical Image Applications, LNCS Vol. 3765, 2005.

[2] Anne E. Carpenter, Thouis R. Jones, Douglas B. Wheeler, David A. Guertin, In-Han Kang, Pollina Golland, and David M. Sabatini. CellProfiler: versatile software for high throughput image analysis. In preparation.

vertical line
vertical line
horizontal line

MIT logo Computer Science and Artificial Intelligence Laboratory (CSAIL)
The Stata Center, Building 32 - 32 Vassar Street - Cambridge, MA 02139 - USA
tel:+1-617-253-0073 - publications@csail.mit.edu