Unité de Bioinformatique génomique et structurale


predict the local structure of a protein in terms of backbone dihedral angle domains, identify sequence regions that form early during folding, and locate structural weaknesses, defined as regions whose sequence is not optimal with respect to the tertiary fold.

These programs use a statistical backbone torsion angle potential, which describes local interactions along the chain and is derived from a set of 1403 known protein structures. The input they require is the amino acid sequence of the target protein. The output yields the predicted structures given in terms of 7 backbone torsion angle assignments noted A, B, C, G, E, O, P; the phi, psi, and omega values associated to these assignments are given here.


predicts the N backbone conformations of lowest energy of a protein sequence or sequence segment, where N is specified by the user. The user may impose constraints on some interatomic distances in the predicted structures. The output file contains the lowest energy structures that satisfy the constraints, the predicted energy values, and the energy gap and root mean square (rms) deviation of superimposed backbone atoms of each predicted structure relative to the lowest energy predicted structure. In addition, the 3D structures of all the predicted conformations are supplied in PDB format.

A sequence whose lowest energy conformation displays a sizable energy gap relative to other predicted structures is considered to have a well defined preferred conformation. The program is designed to run mainly on short peptides, because it only considers local interactions along the chain and overlooks tertiary interactions. When applied to longer peptides, the predictions must be considered as 2D rather than 3D and similar to secondary structure predictions.


predicts the backbone structure, not of the whole input sequence, but of those segments whose lowest energy structure is strongly preferred over other conformations. The strength of the prediction is given by a weight between 1 and 9.

Regions predicted by Fugue, with high weights, have an strong intrinsic preference for their predicted conformation in the absence of tertiary interactions. When cut from the sequence, the peptides so obtained are likely to adopt preferentially this conformation in water at low temperature or in an apolar medium. When considered within the sequence, these regions can be exepected to form at the very beginning of the folding process. Regions where the predicted structure differs from the native structure are likely to be regions whose instrinsic structural preference, determined by local interactions along the chain, are modified due to tertiary interactions. They are interpreted as structural weaknesses, which possibly slow down folding and cause alternative structuring.


  • Prediction of protein backbone conformation based on seven structure assignments. Influence of local interactions.
    Rooman MJ, Kocher JP, Wodak SJ.
    J Mol Biol. 1991 Oct 5;221(3):961-79. Medline
  • Extracting information on folding from the amino acid sequence: accurate predictions for protein regions with preferred conformation in the absence of tertiary interactions.
    Rooman MJ, Kocher JP, Wodak SJ.
    Biochemistry. 1992 Oct 27;31(42):10226-38. Medline
  • Using Prelude&Fugue as a 2D structure prediction method

  • For an assessment of the performances, click here

  • Using Prelude&Fugue to determine peptides that adopt a preferred conformation in solution

  • For an example where Prelude&Fugue were used to propose sequences of helical peptides, which were synthesized and shown by CD and NMR to be indeed helical in a mixture of water and TFE at room temperature, see:
      Conformational properties of four peptides corresponding to alpha-helical regions of Rhodospirillum cytochrome c2 and bovine calcium binding protein.
      Pintar A, Chollet A, Bradshaw C, Chaffotte A, Cadieux C, Rooman MJ, Hallenga K, Knowles J, Goldberg M, Wodak SJ.
      Biochemistry. 1994 Sep 20;33(37):11158-73. Medline

  • Using Prelude&Fugue to identify early folding regions

    For an analysis of the conservation of segments predicted to have well defined instrinsic structural preferences in families of homologous regions, and for a comparison with experimental data on early folding intermediates, see:
      Extracting information on folding from the amino acid sequence: consensus regions with preferred conformation in homologous proteins.
      Rooman MJ, Wodak SJ.
      Biochemistry. 1992 Oct 27;31(42):10239-49. Medline

  • Using Prelude&Fugue to identify structural weaknesses

    Prelude&Fugue have been applied to proteins known to adopt alternative structures that provoke conformational diseases or correspond to 3D domain swapping. In these proteins, sequence regions have been detected whose intrinsic preferred conformations differ from the native conformation, and could represent structural weaknesses that facilitate misfolding. For an application to the amylodoigenic prion proteins, see :
      PoPMuSiC, an algorithm for predicting protein mutant stability changes: application to prion proteins.
      Gilis D, Rooman M.
      Protein Eng. 2000 Dec;13(12):849-56. Medline
    For an application to proteins that undergo 3D domain swapping, see :
      Sequence-structure signals of 3D domain swapping in proteins.
      Dehouck Y, Biot C, Gilis D, Kwasigroch JM, Rooman M.
      J Mol Biol. 2003 Jul 25;330(5):1215-25. Medline