Paper data

In the paper, we used a large set of protein structures and multiple sequence alignments derived for them. In order to allow for validation of results, the entire data archive can be downloaded here (~800 MB) or via BitTorrent (coming soon).

Archive contains following files:
  1. [pdbid].fas - alignment that prediction is based on
  2. [pdbid].pdb - structure of native protein
  3. [pdbid].CB - interresidue distances for Cbeta criterion
  4. [pdbid].heavy - interresidue distances for heavy atom criterion
  5. [pdbid].psicov - predictions by PSICOV
  6. [pdbid].plmdca - predictions by plmDCA (without gap term)
  7. [pdbid].gplmdca - predictions by gplmDCA (subject of the paper)