Fluorescence microscope images of HeLa cells using ten different labels
- DAPI to label DNA
- a monoclonal antibody against an Endoplasmic Reticulum (ER) antigen
- a monoclonal antibody against giantin, a Golgi protein
- a monoclonal antibody against gpp130, a Golgi protein
- a monoclonal antibody against human LAMP2 (primarily found in lysosomes)
- a monoclonal antibody against an outer membrane protein of mitochondria
- a monoclonal antibody against nucleolin
- a monoclonal antibody against transferrin receptor (primarily found in the plasma membranes and endosomes)
- rhodamine-conjugated phalloidin, which labels F-actin
- a monoclonal antibody against beta-tubulin
These images were used in our image classification
and image set comparison
projects, and have been referenced in the following publications:
R. F. Murphy, M. V. Boland and M. Velliste (2000). Towards a
Systematics for Protein Subcellular Location: Quantitative Description
of Protein Localization Patterns and Automated Analysis of Fluorescence
Microscope Images. Proc Int Conf Intell Syst Mol Biol (ISMB 2000) 8: 251-259. [PDF Reprint with high res images (3.8 MB)] [PDF Reprint with compressed images (176 K)]
M. V. Boland and R. F. Murphy (2001). A Neural Network Classifier Capable of Recognizing the Patterns of all Major Subcellular Structures in Fluorescence Microscope Images of HeLa Cells. Bioinformatics 17:1213-1223.
[PDF Reprint] [Screen PDF]
E.J.S. Roques and R.F. Murphy (2002). Objective evaluation of differences in protein subcellular distribution. Traffic 3: 61-65.
[PDF Reprint]
K. Huang and R.F. Murphy (2004). Boosting accuracy of automated
classification of fluorescence microscope images for location proteomics.
BMC Bioinformatics 5:78.
[PDF Reprint]
The images are available as tar archives that have been compressed with
gzip. Each
will expand to 62 MB after extraction and decompression. The images have
been computationally deconvolved using the nearest-neighbor algorithm,
and have been cropped to include one cell per image. All pixels
outside the cropped region have been set to 0.
The SLF4 features used in the ISMB and Bioinformatics papers are also
available. The first line lists the SLF numbers and the second line lists the
short names of the features.
The 180 features used in the BMC Bioinformatics paper listed above to
create feature set SLF16 are
also available as a
Matlab .mat file
(2 MB).