ProMAT

ProMAT takes microarray data (the output of microarray image analysis) where some arrays are treated with standards of known antigen concentration and some arrays are treated with samples of unknown concentration. ProMAT fits standard curves to the standards data to relate spot fluorescence to concentration. ProMAT also calculates and outputs confidence bounds on these standard curves. The tool then uses the standard curves to predict antigen concentrations for the unknown samples, along with prediction intervals.

ProMAT also produces diagnostic plots to aid the user in determining where there may be problems with the data. The figure above shows one such diagnostic plot. The lower right panel shows an example standard curve (in black) with its prediction intervals (in blue) plotted with the standard data. The gray region is the region of the curve where concentration predictions are the "best." The histogram in the lower left panel shows the sample spot values. The upper panel gives the coefficient of variation of the standard curve. The plot quickly shows the user how well the standard curve fits the data, how wide the prediction intervals are, and how well the range of the sample data matches up to the range of the standard curve. This type of plot is produced for each antigen and is displayed in an HTML interface for easy viewing.

The statistical algorithms implemented in ProMAT are described in the following paper:

Daly DS, AM White, SM Varnum, KK Anderson, and RC Zangar. "Evaluating concentration estimation errors in ELISA microarray experiments." BMC Bioinformatics 2005, 6:17. http://www.biomedcentral.com/1471-2105/6/17