Supplementary data for "Large-scale prediction of protein-protein interactions from structures"

Martial Hue, Michael Riffle, Jean-Philippe Vert and William Stafford Noble

BMC Bioinformatics. 11:144, 2010.

Electronic version of the article

---

• Benchmarks

  Each file contains a list of tab-delimited pairs of PDB IDs and their labels, +1 for interacting pairs, -1 for non-interacting pairs.

  • DIP small scale
  • DIP CORE
  • DIP CORE subset
• Matrices

  The matrices are in tab-delimited format. The first row and the first column contain the PDB IDs of the proteins. The MAMMOTH kernel matrix is computed from the MAMMOTH similarity matrix.

  • MAMMOTH similarity matrix (96M)
  • MAMMOTH kernel matrix (320M)
  • mismatch kernel matrix (323M)