1) perform machine learning-based gene filtering with positive sample-only learning algorithm for identifying a set of candidate genes with four different classes of expression characteristics, including the absoulte expression values, the within-condition expression variations, the between-condition expression changes, and the coefficient of variation;
2) construct gene co-expression networks from gene expression data with seven optional methods, including five correlation and two non-correlation measures;
3) perform a comprehensive network comparision with more than thirty network-based characteristics including degree, closeness, eccentricity, eigenvector, and PageRank;
4) identify biologically important genes with different ML algorithms by combining network-based characteristics generated from differential network analysis;
5) estimate the covergence degree between different experimential conditions;
7) quantify the activity of pathways;
8) detect condition specifcally expressed genes.
The tutorial of the mlDNA package can be found at: http://www.cmbb.arizona.edu/mlDNA.
| Package: |
| mlDNA |
| Type: |
| Package |
| Version: |
| 1.1 |
| Date: |
| 2013-11-18 |
| License: |
| GPL(>=2) |