Loading…
Please wait, processing your request…

S8kPred Consensus

Combine multiple prediction tools for a more reliable, bias-reduced secondary structure result

Submit a Consensus Job

Consensus Secondary Structure Prediction

Select two or more tools and provide a FASTA sequence or file

⚠ Please select at least 2 tools for a meaningful consensus.

Pick at least two tools for a meaningful consensus.

About Consensus Secondary Structure Prediction

In consensus secondary structure prediction, outputs from multiple individual prediction methods are combined to generate a more reliable and robust final prediction. Because different algorithms—ranging from statistical models to machine learning (ML) and deep learning (DL)-based approaches—possess distinct strengths and limitations, consensus strategies help reduce individual biases and improve overall predictive accuracy. A well-known implementation of this concept is the NPS@consensus online platform (PRABI Lyon-Gerland, Institute of Biology and Chemistry of Proteins IBCP), which integrates results from prediction programs such as SOPMA, SOPM, PHD, DSC etc., to produce a unified secondary structure assignment (Combet et al., 2000).

In the present consensus prediction approach, we incorporate four widely used and more recent secondary structure prediction tools: PSIPRED v4 (Buchan et al., 2024), Porter 5 (Torrisi et al., 2019), S4Pred (Moffat & Jones, 2021), and S8kPred (Kumar & Rathore, 2026). Such meta-predictor frameworks have proven valuable in enhancing prediction confidence and reliability.

Tools Included

S8kPred v1.0
S8kPred
Tripeptide propensity + PSSM + ML. Predicts Q3 and Q8 states with up to 93% accuracy using 8,000 tripeptide conformational preferences.
Website
PSIPRED v4.0
PSIPRED
Deep learning-based predictor using position-specific scoring matrices. One of the most widely cited tools in secondary structure prediction.
DOI GitHub License: Academic
Porter 5 v5.0.334
Porter 5
Fast ab initio prediction in 3 and 8 classes using cascaded recurrent and convolutional neural networks with evolutionary profiles.
DOI GitHub License: CC BY-NC-SA 4.0
S4Pred v1.2.4
S4Pred
Single-sequence predictor using a deep semi-supervised learning framework, providing high accuracy without needing multiple sequence alignments.
DOI GitHub License: GPL 3.0

References

Combet, C., Blanchet, C., Geourjon, C., & Deléage, G. (2000). NPS@: Network Protein Sequence Analysis. Trends in Biochemical Sciences, 25(3), 147–150.

Rost, B., & Sander, C. (1993). Prediction of protein secondary structure at better than 70% accuracy. Journal of Molecular Biology, 232(2), 584–599.

Jones, D. T. (1999). Protein secondary structure prediction based on position-specific scoring matrices. Journal of Molecular Biology, 292(2), 195–202.

Torrisi, M., Kaleel, M., & Pollastri, G. (2019). Deeper profiles and cascaded recurrent and convolutional neural networks for state-of-the-art protein secondary structure prediction (Porter 5). Scientific Reports, 9, 12374.

Kumar, Mayank & Rathore, R. S. (2026). S8Kpred: a Novel Approach to Protein Secondary Structure Prediction Using 8,000 Tripeptide Propensities. Peptide Science 118(3): e70029. https://doi.org/10.1002/pep2.70029. (S8kPred v1.0: Website)

Buchan, D. W. A., Moffat, L., Lau, A., Kandathil, S. M., & Jones, D. T. (2024). Deep learning for the PSIPRED Protein Analysis Workbench. Nucleic Acids Research, 52(W1), W287–W293. (PSIPRED v4.0) DOI: 10.1093/nar/gkae328

Torrisi, M., Kaleel, M., & Pollastri, G. (2018). Porter 5: fast, state-of-the-art ab initio prediction of protein secondary structure in 3 and 8 classes. bioRxiv. (Porter 5 v5.0.334) DOI: 10.1101/289033

Moffat, L., & Jones, D. T. (2021). Increasing the accuracy of single sequence prediction methods using a deep semi-supervised learning framework. Bioinformatics, 37(21), 3744–3751. (S4Pred v1.2.4) DOI: 10.1093/bioinformatics/btab491