Method automates capture of cell image data

A new type of drug profiling will be useful in identifying the biological targets of experimental compounds and predicting drug toxicity. “This work brings microscopy into the ‘omics’ era,” said Timothy Mitchison, the Hasib Sabbagh professor of systems biology, codirector of the Harvard Medical School Institute of Chemistry and Cell Biology (ICCB) and co-author on the work, which was reported in the Nov. 12, 2004 Science. In the study, the researchers treated human cancer cells with widely varying concentrations of 100 different chemical compounds known to affect cell growth and metabolism. To measure changes in the cells’ behavior, they used fluorescent stains for DNA and 10 different proteins. Automated fluorescent microscopes collected pictures of 70 million cells representing all combinations of compounds, concentrations, and stains. Then, a computer program developed in Mitchison’s lab made 93 different measurements on the cell images, effectively converting the pictures into numbers — lots of them. Then a new number-crunching algorithm written at Harvard’s Bauer Center for Genomics Research organized the mass into a biologist-friendly format.