Consensus-based Secondary Structure predictor (SSCon)

SSCon webserver

Two versions of SScon are available: SScon5 that includes homology modeling-based secondary structure (SS) predictors and SScon3 that consider only ab initio methods. When using SScon5, the user is requested to obtain predictions from the five input SS predictors including SSpro (4.03), PROTEUS, SPINE X, PSIPRED 2.5 and PORTER (each method is linked to its web server). Similarly, to use SScon3, the user should collect prediction from the three ab initio predictors: SPINE X, PSIPRED and PORTER.

Please follow the four steps below to make predictions:

1. Upload file with predicted secondary structure

Up to 1000 proteins are allowed. Protein ID will be trimmed to 10 characters. Please consult help to learn how to format input file.

2. Choose desired predictor

3. Provide your e-mail address (required)

Please provide your e-mail address to be notified when results are ready.

4. Predict:

Click button to launch prediction.

Help

SScon accepts either single or multiple protein sequences. The input is limited to 1000 protein sequences at the time. The user should submit the protein sequence(s) in FASTA format.

The format of the input file for SScon5 is as follows (example):

  1. header that starts with >, which is used to identify a given sequence;
  2. protein sequence using one-letter amino acid code;
  3. predicted SS state from SSpro;
  4. predicted SS state from PROTEUS;
  5. confidence score of PROTEUS prediction separated using commas;
  6. predicted SS state from SPINE X;
  7. probability of strand prediction from SPINE X separated using commas;
  8. probability of coil prediction from SPINE X separated using commas;
  9. probability of helix prediction from SPINE X separated using commas;
  10. predicted SS state from PSIPRED;
  11. probability of strand prediction from PSIPRED separated using commas;
  12. probability of coil prediction from PSIPRED separated using commas;
  13. probability of helix prediction from PSIPRED separated using commas;
  14. predicted SS state from PORTER;
  15. probability of strand prediction from PORTER separated using commas;
  16. probability of coil prediction from PORTER separated using commas;
  17. probability of helix prediction from PORTER separated using commas;

The format of the input file for SScon3 is as follows (example):

  1. header that starts with >, which is used to identify a given sequence;
  2. protein sequence using one-letter amino acid code;
  3. predicted SS state from SPINE X;
  4. probability of strand prediction from SPINE X separated using commas;
  5. probability of coil prediction from SPINE X separated using commas;
  6. probability of helix prediction from SPINE X separated using commas;
  7. predicted SS state from PSIPRED;
  8. probability of strand prediction from PSIPRED separated using commas;
  9. probability of coil prediction from PSIPRED separated using commas;
  10. probability of helix prediction from PSIPRED separated using commas;
  11. predicted SS state from PORTER;
  12. probability of strand prediction from PORTER separated using commas;
  13. probability of coil prediction from PORTER separated using commas;
  14. probability of helix prediction from PORTER separated using commas;

Materials

Datasets:
  • Training dataset in fasta format, available for download at link (training.txt)
  • Test dataset in fasta format, available for download at link (test.txt)

Training set contains 987 proteins; they were used to design the consensus model. Test set includes 988 proteins; they were utilized to perform out-of-sample evaluation and comparison with existing standalone SS predictors. Note that training and test datasets share low (<25%) pairwise sequence similarity.

Standalone

The standalone version of SSCon is available for free, for non commercial users, at this location (link).

Acknowledgments

We acknowledge, with thanks, that the following software was used as a part of this server: