top of page


BioCodeKb - Bioinformatics Knowledgebase

Secondary structure prediction is an important tool in a structural biologist's toolbox for the analysis of the significant numbers of proteins, which have no sequence similarity to proteins of known structure. Jpred is a secondary structure prediction server that is a well-used and accurate source of predicted secondary structure.

The recent update of Jpred incorporates the latest version of the Jnet algorithm improving secondary structure prediction to 81.5% and solvent accessibility predictions to up to 88.9%. The most obvious changes to the server are to do with user interaction with Jpred by giving more feedback to users regarding problems or progress of their submissions.

It has been in operation since approximately 1998. JPred incorporates the Jnet algorithm in order to make more accurate predictions. In addition to protein secondary structure JPred also makes predictions on Coiled-coil regions.


  • Retrained on the latest UniRef90 and SCOPe/ASTRAL version of Jnet (version 2.3.1), mean secondary structure prediction accuracy of >82%

  • Upgraded the Web Server to the latest technologies (Bootstrap framework, JavaScript) and updating the web pages thus, improving the design and usability through implementing responsive technologies

  • Upgraded the results reporting, both, on the web-site, and through the optional email summary reports: improved batch submission, added results summary preview through Jalview results visualization summary in SVG and adding full multiple sequence alignments into the reports

  • Improved help-pages, incorporating tool-tips and adding one-page step-by-step tutorials.


  • New, friendlier user interface

  • Better error checking of input sequences/alignments

  • Predictions now (optionally) returned via e-mail

  • Users may provide their own query names for each submission

  • JPred now makes a prediction even when there are no PSI-BLAST hits to the query

The Jpred server takes a single protein sequence or multiple sequence alignment (MSA) and returns predictions made by the Jnet algorithm. The main differences in Jnet v2.0 are the use of only PSI-BLAST Position-specific scoring matrix (PSSM) and HMMER hidden Markov model (HMM) profiles rather than including frequency profiles, and moving from 9 to 100 hidden units in the neural networks. The method was developed through 7-fold cross-validated training on a sequence and structure non-redundant dataset derived from the Astral compendium of SCOP domain data at the superfamily level. Testing on a blind dataset of 149 sequences gave a final secondary structure prediction Q3 score of 81.5%, which is nearly 5% better than previously published for Jnet.

For a basic search simply input a sequence as a single string and JPred will make a prediction using default parameters. If we wish to change the defaults then use the 'Advanced' options.

Advanced Options

For all searches there is a sequence length limit of 800 residues. Longer sequences should be split into individual domains and searched separately.

Input Sequences

JPred requires input as either a single sequence, cut and pasted into the text box, or a multiple sequence alignment uploaded as a file. It is important not to cut and paste a multiple sequence alignment as the formatting can get messed up. When pasting in a sequence do not add any comment or description lines, but spaces and carriage returns are allowed. We can also upload single sequences as files.

The upload option takes precedence if both inputs have been filled.

Input Format

JPred accepts five types of input. Batch submission of multiple sequences for prediction is via file upload only. The file must be in Fasta format and each sequence must be given a unique name (up to 25 characters with no spaces). Additional words or descriptions on the defline will be ignored.

A limit of 200 sequences are allowed per batch submission. We may submit certain batch jobs, but there is a hard limit of 4,000 sequence predictions in total per user per day.

Searching the PDB

By default JPred searches the PDB before doing a prediction. If our sequence has homology to a known structure we should reconsider the utility of making a prediction: a prediction is unlikely to be better than using the structure of a known homologue.

E-mail (Optional)

Providing an e-mail address means we will be notified of the fate of our job. A simple summary of the result will be emailed to us when the prediction is complete.


Need to learn more about Jpred and much more?

To learn Bioinformatics, analysis, tools, biological databases, Computational Biology, Bioinformatics Programming in Python & R through interactive video courses and tutorials, Join BioCode.

bottom of page