• Prediction of protein secondary structures
  • DNAMAN predicts the secondary structure of the current protein sequence using the DSC (Discrimination of protein Structure Classes) method developed by King and Sternberg (Protein Science, 1996, 5: 2298-2310). This method can provide relative accurate and reliable protein secondary structure prediction.

    1. Choose the Protein | Secondary Structure command to start prediction for the current protein sequence.

    2. A dialog box appears. There are two essential parameters.

    - Filter Level. This parameter is used to filter the prediction according to prediction rules. The default level is 1.

    - Remove singlet option. This option is checked by default. In this case, if a residue has different structure from its neighbors, it will be changed according to its neighbors’ structure.

    3. If you are predicting the secondary structure of multiple alignment protein sequences, there are three parameters to set.

    - Remove poor aligned regions option. This option is checked by default. In this case, DNAMAN will ignore the poor aligned regions for prediction. Poor region is defined as the two following parameter.

    - Length of poor region. The default is 40 residues.

    - Homology <(%). The default is 20%. If a region with a defined Length has less than this percentage, DNAMAN will considerate this region as poor aligned region.

    4. Click OK to display the results in a text window and a graphic window.

    The prediction of protein secondary structures is shown as a table in a text window. You may save the table and plot it with other Graphic/Table programs, such as Microsoft Excel. For the amino acid at a given position, DNAMAN displays the probability of forming alpha-helix, beta-sheet, or random coil. The most possible structure will be assigned to this residue. An overall prediction accuracy will be calculated.

    DNAMAN also presents the prediction in a graphic window. The structure probabilities are plotted against the positions. See Chapter V for more information about the handling of Profile Plot.