TM4 Microarray Software Suite

Category Genomics>Gene Expression Analysis/Profiling/Tools

Abstract The TM4 Microarray Software Suite of tools consist of four (4) major applications: 1) Microarray Data Manager (MADAM); 2) TIGR Spotfinder; 3) Microarray Data Analysis System (MIDAS); and 4) Multiexperiment Viewer (MeV), as well as a Minimal Information About a Microarray Experiment (MIAME)-compliant MySQL database, all of which are freely available to the scientific research community at TIGR's Software Download Site.

Although these software tools were developed for spotted two-color arrays, many of the components can be easily adapted to work with single-color formats such as filter arrays and GeneChips™ (Affymetrix). Three (3) of the TM4 applications, MADAM, MIDAS, and MeV, were developed in Java; TIGR Spotfinder was written in C/C++.

1) The Microarray Data Manager (MADAM) -- facilitates the entry of data into a relational database. MADAM guides users through the microarray process from RNA procurement to data analysis, offering intelligent forms to simplify the tracking of experimental parameters and results that are essential for the interpretation of expression results in downstream analyses.

Canned reports provide information on RNA samples, studies, slide maps and other pertinent data and a general SQL query window allows freeform access to the underlying database. MADAM also serves as a platform for launching other data entry and management tools.

Through the use of these integrated modules, users can view and score PCR plates, design experiments and studies, and track laboratory materials.

2) TIGR Spotfinder -- Image analysis is a crucial step in the microarray process. TIGR Spotfinder was designed for the rapid, reproducible and computer-aided analysis of microarray images and the quantification of gene expression. TIGR Spotfinder reads paired 16-bit or 8-bit TIFF image files generated by most microarray scanners.

Semi-automatic grid construction defines the areas of the slide where spots are expected. Automatic and manual grid adjustments help to ensure that each rectangular grid cell is centered on a spot. Two (2) available segmentation methods (histogram and Otsu) define the boundaries between each spot and the surrounding local background.

Spot intensities are calculated as an integral of non-saturated pixels, although other options including spot median and mean values are available.

Local background subtraction for each reported value is applied by default but can be disabled. The calculated intensities, medians, and means along with each spot position on the array, spot area, background values, and quality control flags are written to a MeV file or the database.

Reusable grid geometry files and automatic grid adjustment allow user to analyze large quantities of images in a consistent and efficient manner. To complement the automated methods, particularly in noisy areas of the slide, the user may manually identify or discard spots. Quality control views allow the user to assess systematic biases in the data.

3) Microarray Data Analysis System (MIDAS) -- Before the intensity values measured in TIGR Spotfinder can be compared, normalization is necessary. This critical step can help compensate for variability between slides and fluorescent dyes, as well as other systematic sources of error, by appropriately adjusting the measured array intensities.

Data filtering can reduce the dataset by removing poor or questionable data, in addition to data deemed uninteresting or irrelevant to the analysis.

MIDAS provides users an intuitive interface to design analysis protocols combining one or more normalization and filtering steps. In this way, data from many individual hybridizations can be treated in a uniform and reproducible manner. MIDAS reads “.tav” files generated by 'TIGR Spotfinder' or retrieved from the database via MADAM.

Normalization modules include locally weighted linear regression [lowess; (Cleveland and Devlin, 1988; Yang et al., 2002)] and total intensity normalization. These can be linked with filters, including low- intensity cutoff, intensity-dependent Z-score cutoffs, and replicate consistency trimming, creating a highly customizable method for preparing expression data for subsequent comparison and analysis.

Data analysis methods are constructed using an intuitive graphical scripting language and can be saved for application to other datasets. MIDAS provides scatter-plots that illustrate the effects of each algorithm on the data. When the normalization and filtering steps are complete, MIDAS outputs the data in tav format.

4) Multiexperiment Viewer (MeV) -- Normalized and filtered expression files can be analyzed using MeV. MeV is a microarray data analysis tool, incorporating algorithms for clustering, visualization, classification, statistical analysis and biological theme discovery.

MeV can handle several input file formats. These include the “.mev” and “.tav” files generated by TIGR Spotfinder and TIGR MIDAS, and also Affymetrix® (“.txt”) and Genepix® (“.gpr”) files. MeV generates informative and interrelated displays of expression and annotation data from single or multiple experiments.

At this final stage of the TM4 pipeline, flexibility and the variety of analysis techniques are critical, as every algorithm has strengths that can be exploited when used on certain datasets and experimental designs.

The concept of modularization lends itself particularly well to this system, as novel algorithms and existing code bases from the microarray community can be integrated with the Java based MeV using a well-defined module Application Programming Interface (API).

Algorithms such as, bootstrapping, jackknifing and k-means support, resample the dataset to generate consensus clusters. Figure of merit graphs (FOM; (Yeung et al., 2001) suggest appropriate input parameters for algorithms such as k-means.

Clusters identified through any analysis method can be labeled and tracked through other analyses, providing the user with the ability to compare the results of several clustering algorithms to determine consensus and focus on genes with specified expression patterns and biological profiles.

System Requirements

TM4 requires JAVA v1.4.1 or higher. Spotfinder, MIDAS and MeV require Windows 2000/XP, MacOSX or Linux. MADAM currently only runs on Windows 2000/XP. Hardware requirements for all of these packages depend on the size of data sets used.


Manufacturer Web Site TM4 Microarray Software Suite

Price TM4 is free, open-source software released under the Artistic license. This license is OSI certified.

G6G Abstract Number 20224

G6G Manufacturer Number 100707