Human HT29 Colon Cancer 1
![]()
Biological application
These images are of human HT29 colon cancer cells, a cell line that has been widely used for the study of many normal and neoplastic processes. A set of about 43,000 such images used by Moffat et al. (Cell, 2006) to screen for mitotic regulators. The analysis followed the common pattern of identifying and counting cells with a phenotype of interest (in this case, cells that were in mitosis), then normalizing the count by dividing by the total number of cells. Such experiments depend on accurate cell counts.
Images
The image set consists of 6 images. The samples were stained with Hoechst 33342, pH3, and phalloidin. Hoechst 33342 is a DNA stain that labels the nucleus. Only the images of the Hoechst 33342 channel are included, as the other channels are not useful for counting cells.
The cells in all six images have been treated with an RNA interference reagent (Rock1_1885_k27) in the course of a high-throughput screen; see Moffat et al. (Cell, 2006) for detail. The images were acquired at the Whitehead-MIT Bioimaging Center on a Cellomics ArrayScan. The images provided here are a single channel, DNA. Image size is 512 x 512 pixels. Images are available in native DIB format or 8-bit TIF.
Download DIB images (ZIP) | Download TIF images (ZIP)
Ground truth 
A tab-delimited text file contains cell counts in each of the 6 images, as determined by two different human counters. To compare an algorithm's results to these, first compute for each image the absolute difference between the algorithm's count and the average of the humans' counts, then divide by the latter to obtain the deviation from ground truth (in percent). The mean of these values over all 6 images is the final result.
The two human observers vary by 11% for this image set.
Published results using this image set
| 6.2% | Carpenter et al., Genome Biology, 2006. See Figure 2A, pages 6-7. Example CellProfiler pipeline for counting human cells available at www.cellprofiler.org/examples.htm. |
Recommended citation
"We used the Human HT29 Colon Cancer 1 image set (Carpenter et al., Genome Biology, 2006) available from the Broad Bioimage Benchmark Collection (www.broad.mit.edu/bbbc)."





