Previous - Next - Bottom - PP home - PP help TOC

PP Help 05: Advanced prediction options

Contents

Advanced prediction options

  1. PHDsec only
  2. PHDacc only
  3. PHDhtm only
  4. toggle filtering ali for PHD
  5. toggle ProSite reports
  6. toggle SEG assignments
  7. toggle ProDom assignments
  8. toggle COILS prediction
  9. toggle CYSPRED prediction
  10. toggle ASP prediction
  11. TOPITS threading
  12. Evaluate prediction accuracy (for developers)


ADVANCED PREDICTION OPTIONS

The following prediction options have been implemented:

  1. Predict only secondary structure ('predict secondary structure')
    The string "predict secondary structure" (or short: "PHDsec") in any line before the one beginning with a hash (#) results in a prediction of secondary structure, only.
  2. Predict only solvent accessibility ('predict accessibility')
    The string "predict accessibility" (or short: "PHDacc") in any line before the one beginning with a hash (#) results in a prediction of solvent accessibility, only.
  3. Predict only transmembrane helices ('predict transmembrane')
    The string "predict transmembrane" (or short: "PHDhtm") in any line before the one beginning with a hash (#) results in a refined prediction of transmembrane helices and topology.

  4. Filtering PHD input alignment (Default)
    If the divergence found in your family is not 'well' spread, prediction accuracy may drop. In particular, too many highly similar sequences may be problematic in absence of further diverged family members. This problem came up only in the post-genome era, i.e. since the number of sequences is exploding. To correct for this problem we run a crude filter on the alignment, by default. To switch this filter off, please use the keyword 'no filter' in any line before the one commencing with a hash ('#').

  5. ProSite sequence motifs (Default, example)
    ProSite sequence motifs are reported by default, when found. To surppress this default, use the string "no prosite" in any line before the one beginning with a hash (#).
  6. Low-complexity regions by SEG (Default, example): The program SEG (J C Wootton & S Federhen) is executed, by default. SEG scans your sequence for regions of low-complexity ('simple sequences' or 'composition-biased regions'). You may turn this default off by using the keyword 'no seg in any line before the one beginning with a hash ('#').
  7. ProDom domains (Default, example)
    ProDom domains are reported by default, when found. To surppress this default, use the string "no prodom" in any line before the one beginning with a hash (#).
  8. Prediction of cysteine bridges (Default, example)
    Predicted cysteine-bridges are reported by default, when found. To surppress this default, use the string "no cyspred" in any line before the one beginning with a hash (#).
  9. Prediction of structural switching regions (Default, example)
    Regions predicted to undergo structural rearrangements are reported by default, when found. To surppress this default, use the string "no asp" in any line before the one beginning with a hash (#).

  10. Prediction-based threading ('prediction-based threading')
    The string "prediction-based threading" (or short: TOPITS) in any line before the one beginning with a hash (#) requests the prediction-based threading (TOPITS). If you do not provide 'your prediction' for the fold recognition, we return a multiple sequence alignment, a prediction of secondary structure and accessibility, and the threading alignment of your sequence and possible remote homologues of known structure.
    The results of the search for remote homologues are given in three blocks. (1) An MSF formatted (default) alignment between your sequence and possible remote homologues. (2) A summary of some statistics such as alignment scores for a cross-validation experiment and for your request. (3) An alignment presenting the entire motifs of 1D structure for the alignment.

  11. Evaluation of secondary structure prediction accuracy (for program developers, only!)
    The output consists of two parts: (1) per-residue and per-segment scores for each protein in your input file, and (2) per-residue and per-segment scores for all proteins in your input file (the latter includes scores for the accuracy in predicting secondary structure content and secondary structural class; example for output).




Previous - Next - Top - PP home - PP help TOC