Genowiz™ is a powerful gene expression analysis program that has been designed to store, process and visualize gene expression data efficiently. It includes a suite of advanced analysis methods and allows researchers to select analysis methods appropriate for their dataset. Genowiz™ allows researchers to organize experimental information (MIAME), import data files quickly and easily, work with multiple experiments at the same time, import gene annotation files, preprocess and normalize data, perform cluster analysis, classify and view gene information, perform functional classification and track down intricate correlations in data by performing pathway analysis. All analysis done is tracked, saved into a database and can be retrieved at any point of time.
Data and Gene List Import
Genowiz™ supports a wide range of data formats pertaining to cDNA data and Affymetrix processed data. Users also have an option to upload data in customized formats. Customized uploader allows users to add and save new data formats. One-Click Uploader can then identify these formats.
Gene List files for annotating genes can also be imported.
Minimum Information About a Microarray Experiment (MIAME) facilitates adoption of standards for microarray experiment annotation and data representation. Genowiz™ focuses on establishing standard microarray experimental data repositories and information sharing within the scientific community. Researchers can also exchange MIAME data by using MAGE ML document exchange format.
Data Transformation, Normalization and Filtration
In any type of expression analysis, pre-processing of data to reduce undesirable variation among datasets and to bring data to a common platform is a vital step. Genowiz™ provides users with a wide range of data transformation, normalization and filtration tools. These include:
• Data transformation options such as imputation of missing values, log transformations, mean/median, Z-transformation, subtract control, divide by control, scaling etc.
• Normalization techniques such as normalization for dye swap replicates, cDNA raw data normalization options (cDNA Loess and Print tip Loess) and quantile normalization. Separate normalization techniques are provided for cDNA and Affymetrix arrays. Normalization can be done using all genes or control genes.
• Filter data based on replicate genes, fold change, mean, standard deviation, calls and missing values. Replicate samples are handled using various parametric/non-parametric tests. Multiple testing correction can be applied to reduce false positives.
Data Analysis and Visualization
Genowiz™ comes equipped with several data analysis tools. Complete with excellent graphics, it is an excellent tool for interpretation of biologically meaningful results. Some of these tools include partition clustering, hierarchical clustering, SOM, PCA, gene shaving and discriminant PCA (for classification).
• Partition Clustering (k-means, Forgy's)
This tool classifies genes or samples in user-defined groups using distance parameters. The obtained clusters can be re-clustered. Re-clustering utility helps scientists pick a set of genes of their interest. A 2D PCA view shows the distribution of genes in various clusters.
• Hierarchical Clustering
One of the most important tools for studying relations between genes, this tool creates a dendrogram based on the relative distance between genes. The different optional parameters help the user in correctly determining the relationship between two genes. Models of analysis include single linkage, complete linkage and average linkage clustering. Genes, samples, or both together can be clustered.
• Self Organizing Maps
A two-way classification of genes into clusters based on novel artificial neural networks is an integral feature of data clustering tools in Genowiz™. This gives a deeper insight into clusters, as neighboring clusters are very similar to each other.
• Principal Component Analysis
This tool involves a mathematical procedure that transforms a number of (possibly) correlated variables into a (smaller) number of uncorrelated variables called principal components. These provide an insight into existent variability in the data.
• Gene Shaving
This method identifies subsets of genes with coherent expression patterns and large variation across conditions. Gene shaving differs from hierarchical clustering and other methods of gene expression analysis in that genes may belong to more than one cluster.
• Discriminant PCA
This statistical approach classifies samples of unknown classes based on training samples of known classes.
Genowiz™ annotates genes and classifies them into functional categories (Gene Ontology). Option of importing annotation files is also provided. Integrated pathways module aids researchers in understanding metabolic pathways in relation to expression data. Pathway maps edited/created can be associated with author details too. Coupled with biological information and gene ontological classification, it forms an excellent tool in understanding biological systems.
Several utility options are present adding value to analysis performed:
Gene List Comparison: Subtle relations among datasets can be probed using this feature.
Pattern Simulation: An expression pattern can be defined and Genowiz™ lists out all genes with a similar expression pattern. This gene list can be saved and exported.
Gene Tracking: Important genes or genes of interest can be tagged and tracked throughout the analysis.