Anne Carpenter’s task was not as straightforward as it sounded. She was trying to measure the size of Drosophila fruit fly cells, but she was continually frustrated by the bottleneck in the processing of cell images.

It was the mid 2000s, and the problem was that commercial software for doing high- throughput image processing “was choking on the type of images we had created and doing a terrible job of quantifying cellular features,” says Carpenter, who received her PhD in cell and developmental biology from Illinois in 2003 (under the name Anne Nye). “The Drosophila cells are not as nicely shaped or as uniform as mammalian cells.”

So she took the matter into her own hands. Carpenter decided to write her own software code to solve the imaging problems she faced at the time while doing postdoctoral work at the Whitehead Institute for Biomedical Research in Massachusetts. The result eventually became CellProfiler, an extremely popular and successful open-source software program that biologists around the world now use.

CellProfiler is launched more than 125,000 times each year, and it has been cited in more than 3,000 scientific papers published since the software was released in 2005. The tool can be used for all types of imaging work, such as correcting illumination patterns; identifying cells, yeast colonies, and other biological objects; or measuring features such as the size and quantity of specific cells.

Since 2007, Carpenter has been director of the Imaging Platform at the Broad Institute, a joint venture between Harvard and MIT. In addition to offering the use of CellProfiler to biologists all over the world, her lab collaborates directly with many researchers using the software.

For instance, Carpenter’s lab collaborated with John Crispino from Northwestern University, who recently found a promising drug treatment for AMKL leukemia—a cancer that primarily affects children.

Carpenter used CellProfiler to measure the DNA content of individual cells, which was key to the discovery.

Crispino’s lab identified a drug that causes the leukemic cells to become polyploid. This results in the cells having excess DNA, which the body recognizes as abnormal. The body then uses its own mechanisms to destroy the cancer cells. 

A clinical trial opened last October to test the drug in AMKL patients.

Carpenter’s venture into imaging technology actually began when she was a PhD student in CDB professor Andrew Belmont’s laboratory at Illinois. The lab was interested in figuring out whether particular transcription factors caused chromatin to unfold or not, but the technology available back then made the work “incredibly tedious,” she says.

Belmont had just acquired a new robotic microscope, so Carpenter spent a summer programming the equipment to do the image analysis automatically, cutting out the tedium and speeding the process.

“My experience in Andy Belmont’s lab was a turning point for me,” Carpenter says, “because that’s when I became interested in developing systems to automate biology.

“If you had told me in high school or even college that eventually I would be leading a group of software engineers and computer scientists at Harvard and MIT, I would be utterly confused and surprised,” Carpenter adds. She didn’t even know she would be going into a scientific field until her first year at Wheaton College in the Chicago area. Growing up on a farm in northwestern Indiana, she was a bookworm with little interest in science. In fact, when her grandfather gave her a science kit, she never used it once, and by deferring to lab partners, she went through high school and college avoiding the dissection of anything larger than a fruit fly. But during her first year at Wheaton College, she enjoyed her science classes so much that she transferred to Purdue to get a bachelor’s degree in biology.

Imaging and microscopy have come a long way since those days. Before the 1990s, she says biologists didn’t often capture and measure images; they looked through their microscopes and judged the characteristics of cells subjectively by eye.

In the 90s, things began to change, as the digital capture of cell images emerged. Then, in the 2000s, during her PhD and postdoc years, automated image analysis tools exploded on the scene. So she has ridden the wave of technological change in image analysis from the beginning, and CellProfiler has been part of that transformation.

CellProfiler has been used in research on all sorts of diseases, including Ebola, tuberculosis, and cancer. In another leukemia research project, Carpenter’s laboratory is working with Todd Golub, also with the Broad Institute, in targeting cancer stem cells.

“In many cases of cancer, it’s possible to kill the tumor cells, but it’s very hard to kill the last tiny percentage of the cells—the cancer stem cells,” she says. These stem cells often trigger the recurrence of cancer in patients.

In this project, researchers succeeded in finding a drug candidate that could kill the leukemic stem cells but leave the bone marrow stem cells safely intact. Carpenter’s lab used CellProfiler to distinguish between leukemic stem cells, which have a cobblestone appearance, and bone marrow cells, which are more rounded.

Over the past 10 years, her lab has amassed a huge collection of images, but she says there remains a lot of information that can be extracted from them. Every single cell has thousands of features that can be measured. Therefore, one branch of her lab is harvesting this information from the cell images in their collection.

She says there are a myriad of uses for this kind of information. For instance, one possibility is to create a profile, or fingerprint, of cells’ responses to treatments with various compounds. Researchers can then use this profile to screen compounds for toxicity to cells in the liver, heart, or other organs.

Since the time that Carpenter first created the CellProfiler software in the mid-2000s, the program has been rewritten and improved by her laboratory’s professional software engineers. But one reason CellProfiler has been so successful, she says, is that it “was originally designed from the ground up by a cell biologist who was trying to accomplish something in a real project. It started out as a bit of code that I needed for my project, but grew into a multifunctional toolbox that is changing the world.”

Carpenter received the NSF CAREER award in 2012, and has received recognition and research funding from numerous other groups including the NIH, the Human Frontiers in Science program, and the Howard Hughes Medical Institute. She was namedYoung Leader of the French-American Foundation, and elected fellow of the Massachusetts Academy of Sciences. She was featured in a public television special, “Bold Visions: Women in Science & Technology,” and named a “Rising Young Investigator” by Genome Technology magazine. CellProfiler was awarded the Best Practices Award by Bio-IT World in 2009.