There exists four classes of prediction according to their intrinsic target.
Each one could have more than one predictor, according to the computation
The four classes are:
- Transmembrane Strand Predictors
- Transmembrane Helix Predictors
- Secondary Structure Predictor
- Signal Peptide Predictors
The input data is the following:
- Prediction: it's possible to select only one predictor at the time. Making
multiple prediction is possible both serially and in parallel, but
not from the web interface. In fact, it's possible to resubmit the
same sequence at the end of each prediction, taking advantage of
the form available in the results page (Output page), or opening
new input forms and submitting the sequence to other predictors
(and starting parallel tasks).
- Input: a protein sequence supplied in the standard IUPAC protein alphabet.
Only the twenty canonical amino acids will be accepted and any other
character will be cut out and excluded from the prediction.
- Database: when a database is selected (if available), the predictor
generates a "sequence profile", that usually is more meaningful than
the sequence itself in the detection of structural-functional features.
The default database is the one used to generate profiles for the
training process of each prediction method.
- Submission: in the input sequence space, the "OK" button is used to
submit the sequence to the prediction server. Next, a waiting page
will be prompted until the results are generated.
- Reset: the "Reset" button is used to clear the input sequence space.
The upper menu is as follow:
- Top Menu:
It is a link to the session page that shows old stored predictions
sorted by submission date. The list contains also the session number
and the predictor name.