Software Standalone applications GaMRed

GaMRed

Authors: Roman Jaksik, Michał Marczyk, Andrzej Polański, Joanna Polańska

GaMRed is designed to estimate the noise level of large numerical datasets using Gaussian mixture models. All of the features (rows) of the dataset are automatically removed if they fall below the estimated noise level. This approach can be used to reduce the number of features in microarray or RNA-Seq experiment data, although the method is applicable to any dataset where large number of features was measured for any number of samples.  

This approach can be used to reduce the number of features in microarray or RNA-Seq experiment data, although the method is applicable to any dataset where large number of features was measured for any number of samples.

The application utilizes 4 distinct filtering methods:

• S - This filter is based on the average value of each feature (rows) across all samples (columns) in log2 scale. It is recommended for datasets in which expected differentially expressed genes have high signal.
• LV - This filter is based on the variance of all features (rows) across all samples (columns) in log2 scale. It is recommended for datasets with small number of differentially expressed genes.
• V - This filter is based on the variance of all features (rows) across all samples (columns) in linear scale. This is the most universal filter applicable to most datasets.
• IS – This filter is based on distributions of individual samples which are modeled separately. The final noise level is the minimal value selected from those obtained for individual samples.

For more information please read the software manual.

Available application versions: