Part A - Sequence and Choices

The first screen is composed of a window to insert the studied sequence (1), three buttons to control the software (2-4), three choices to control the type of predictions, eight for the outputs (5,6) and one small text to see the advance of the computation (7).

window sequence explication

1. sequence

amino acid sequence: copy the amino acid sequence in Fasta format.
NB: Fasta format is a simple format : (1) the first line is the name of the protein beginning with a symbol ">" given ">name". (2) The following lines are used for the amino acid sequence. You could use the example button to see the example of protein 2aak.

Top

2. Submission

button submit: When your sequence is written with your prediction and output choices, just push this button.
The "state" will give you the state of the computation.

Top

3. Clean

clean button: If you want to change your sequence, just push this button and everything in the window sequence will be erased.

Top

4. Example

protein 2 aak: This protein was used in our paper as example (de Brevern AG., Etchebest C., and Hazout, S, Proteins, 2000). It is an alpha-beta protein. You can download the 3D structure from the Protein Databank.

Top

5. Prediction

type of predictions: "Bayesian Prediction": the basic prediction is performed with one occurrence matrix per Protein Block.; "Sequence Families" : a Bayesian prediction is performed with optimized matrices.; "Sequence Families (new)" : the new approach has improved the prediction rate from 40.7% to 48.7%.
Each Protein Block could be characterized by several occurrence matrices, each one is associated with a cluster of "related" amino acid sequences (increase the prediction accuracy for 95% to 99.8% of the proteins).

Top

6. Outputs

Displaying of the results: results; "Short results" : only the sequence and the prediction in terms of Protein Blocks are given.; "Results": normal results are the same results than the previous item, but completed with an index derived from the Neq to evaluate the prediction quality.; "Complete results": the complete results are the amino acid sequence, the prediction in terms of Protein Blocks and the whole Protein Blocks with their associated probabilities.; Neq information; "Neq (Local)" : local Neq for every sites of the protein, with the possibility to perform the two different strategies (local strategy and global strategy) for a given prediction rate.; "Neq (Global)" : global Neq for the complete protein, with a rasmol script to visualize the informativity of the sequence directly on the structure (if the 3D structure is available, click here
for more details).; "All Neq Graphics" : The two last options.; "Comparison [L]" : comparison between the prediction in terms of PBs and a 3D structure coded in terms of PBs.; "Score visualisation" : The visualisation of the prediction scores (colored matrix).; "All" : the two Neq graphics and the complete results.

Top

7. state

nothing special: "State of LocPred : sleep " : no submission done.

a wrong sequence: "State of LocPred : no correct sequence " : too many errors in the sequence. So the prediction could not be performed.

too long sequence: "Warning : sequence length too long" : only the beginning of the sequence is predicted. Please cut the sequence into successive fragments. If you want a prediction for longer sequence, cut your amino acid sequence into successive fragments (with overlaps). The maximum length allowed is 2000 residues (mail me if you want a higher value by default).

reading of the sequence: "Sequence read" : no special problem to read the amino acid sequence.

computation of Bayesian prediction: "Bayesian Prediction in progress" : work in progress.; "Bayesian Prediction done" : computation done.

computation of Bayesian prediction with sequence families: "Bayesian Prediction (Sequence Families) in progress" : work in progress.; "Bayesian Prediction (Sequence Families) done" : computation done.

results: "Short Results in progress" : frame for the short results option is in progress.; "Results in progress" : frame for the (second) results option is in progress.; "Complete Results in progress" : frame for the complete results option is in progress.; "Computation done" : the work is done.

Neq: "Neq (Local) in progress" : computation of local Neq is in progress.; "Neq (Global) in progress" : computation of global Neq is in progress.; "Computation done" : the work is done.
NB : The computation is always done in the same order (1) the results, (2) the local Neq, (3) the global Neq, (4) the comparison form and (5) the visualisation matrix.

error messages: If you see one please mail me (it's the message).

end: "Finish" : everything is done. I hope our approach will help you, good luck.

Top

back
Last modif : 11 April 2004