POST

Introduction

Recognizing the oligomeric state of proteins is crucial for understanding structure and function of proteins. It is challenging to accurately predict the oligomeric state of proteins. We introduce POST, a new approach to the prediction of oligomeric state for homo-oligomers using multiple templates. To detect oligomeric templates from an in-house template library of quaternary structures, three different algorithms are used, including dynamic programming, protein language model, and hidden Markov model. They lead to three individual methods to oligomeric state prediction. Assessment on an independent dataset of 1200 proteins, and 141 targets from CASP14 and CASP15 suggests that the templates detected by these methods are largely complementary. Combination of the templates by all individual methods results in the most accurate prediction. For example, POST’s F1-scores in all tested targets are higher than 0.7. As a proof of concept, we show that the predicted oligomeric state can be used together with AlphaFold-Multimer to improve protein structure prediction. POST is anticipated to be helpful in protein structure prediction and protein design.

Submit
  • Provide the protein data (mandatory)
    •
    Input protein data below: (sequence in FASTA format).   (Click for an example input)
    Or upload the sequence file:

  • Other information (optional)
    • Email: (Optional, where the results will be sent to)

    Target name: (Optional, your given name to this target)



    Reference

  • Yuxian Luo, Haiyan Wu, Hong Wei, Jianyi Yang, Zhenling Peng, Predicting the oligomeric state of proteins using multiple template detected by complementary alignment methods, submitted, 2024.