Introduction


Spotfinder is an image processing program created at The Institute for Genomic Research (TIGR) for analysis of the image files generated in microarray expression studies. Spotfinder uses a fast and reproducible algorithm to identify the spots on array and provide quantitation of expression levels.

The Spotfinder 3.0 has multichannel functionality, i.e. it can analyze microarray slides labeled by more then two dyes. The hardcoded maximum channel number is four but can be changed by request. As input Spotfinder expects up to four separate 16 bit or 8 bit gray scale TIFF image files, generated by a confocal laser scanner. Each file is scanned for one fluorescence dye labeling the query or reference samples used in most microarray experiments. Typically, the spots on array are divided into contiguous sections or subarrays or grids. Spotfinder builds set of grids for whole slide in user assisted mode and then process whole slide in one pass. The user is required to enter a number of parameters to build the grid: the spot spacing, number of printing pins in vertical and horizontal dimensions, distance between them, number of columns and rows in each grid. User needs to modify and align the left most top grid in array manually and then propagate alignment to the rest of the grids automatically. During the process pass the program automatically searches each grid element to identify the spots, measures the signal and local background. Output data can be exported to tab-delimited MEV file for whole slide. The MEV file can be passed to other TM4 programs such as TIGR MultiExperiment Viewer (MeV), or MIDAS.

Program limitations.

The use of Spotfinder is limited to the microarray image types and problems we identified and targeted during development. The software requires multiple images generated by scanning a single slide in both the reference(s) and query(s) channels using identical definitions of the scan area. The sensitivity and gain settings can be different for the all channels. Spotfinder uses data from all channels (all loaded images) to define spot boundaries. Non-overlapping images or those of different sizes will cause the program to fail.