Y.Yamanishi, J.-P.Vert, and M.Kanehisa,
Protein Network Inference from Multiple Genomic Data: A Supervised Approach

Predicted protein network for all the proteins of the budding yeast
(6059 ORFs of Saccharomyces cerevisiae)

1. Correlation coefficient matrix between the proteins in feature space (text file)
   
   • Correlation coefficient matrix
   
   
2. Examples of adjacency matrix between the proteins (text file)
   
   • Adjacency matrix when the threshold is 0.7
   • Adjacency matrix when the threshold is 0.5
   • Adjacency matrix when the threshold is 0.3
   
   Note that the numbers of edges can be determined by the threshold of graphical assoications estimated.
   
   

Examples of kernel matrices used in this study
(679 x 679 matrices)

1. Kernel matrix of gold standard protein interactions
   
   • Kmat_gold.txt
   
   
2. Kernel matrix of gene expression
   
   • Kmat_exp.txt
   
   
3. Kernel matrix of yeast two hybrid
   
   • Kmat_y2h.txt
   
   
4. Kernel matrix of localization
   
   • Kmat_loc.txt
   
   
5. Kernel matrix of phylogenetic profiles
   
   • Kmat_phy.txt
   
   
6. Kernel matrix integrating all the genomic data
   
   • Kmat_int.txt
   
   Note that the performance depends on the choice of kernel functions and their parameters.
   
   

Data used in this study

1. Adjacency matrix of proteins in KEGG/PATHWAY
   
   • admat.txt
   
   Note that this is based on the xml files in KEGG.
   The pathway data in the KEGG database are being updated day by
   day, but the data here will not be updated.
   We suggest to see the latest version.