This page contains links to the homology modelling results and downloadable coordinates of the acid-labile subunit protein as generated by the Phyre2 protein structure prediction server.
Phyre2 has two primary modes for model building: Normal and intensive. Below are links to the results of running the ALS sequence through each of these two Phyre2 modelling protocols.
Normal Mode
For 'normal' mode, Phyre2 builds a hidden markov model (HMM) of the user sequence and compares this to a library of HMMs built for a representative set of known protein structures. Regions of the user sequence with no significant match to known structures are simply left unmodelled. In the results below, it can be seen that Phyre2 has modelled residues 43-598 of the ALS protein based on the top scoring homologous template with PDB identifier 3fxi.
Phyre2 results for ALS in 'normal' mode
Intensive Mode
For 'intensive mode' Phyre2 performs the same modelling as normal mode, but additionally attempts to model unmatched stretches of sequence using a coarse-grained molecular dynamics approach called Poing [1]. Because this approach does not rely on homology to known structures for such regions, their predicted structure must be treated with caution. Also, 'intensive' mode attempts to use multiple homologous templates to cover more of the user sequence. In this case Phyre2 used 6 different templates for this model, shown at the bottom of the results page.
Phyre2 results for ALS in 'intensive' mode
Help on interpreting and navigating Phyre2 results pages and detailed information on the method can be found here
A note on confidence measures
Phyre2, HHpred and I-TASSER are well established powerful methods for detecting remote homologous relationships and model building. HHpred and Phyre2 report probabilities for a homologous relationship. This probability is based on the real-world score distribution for negative and homologous domain pairs in an all-against-all comparison of SCOP, the expert curated database of protein domains. Both servers return confidence estimates >95% indicating a very high probability that the ALS sequence is homologous to the templates reported. Confident homology is a very strong indicator that proteins adopt the same or highly similar folds. At the level of sequence identity reported here (20-25%) deviations of the model from the native are expected for structural details, but the overall fold is expected to be correct.
I-TASSER uses a different confidence estimation scheme, calculated based on the significance of threading template alignments and the convergence parameters of the structure assembly simulations. Here we used I-TASSER as an independent validation of the general fold of the ALS model. As can be seen in the figures and viewer below, the I-TASSER model adopts the same fold as that predicted by HHpred and Phyre2 and indeed is based on one of the same templates used by Phyre2 (PDB code 1ZIW).
Supplementary Figures
Below are three figures detailing the level of similarity between the models of ALS generated by the three public servers, and the similarity of the Phyre2 model to the six input templates. Finally, at the bottom of the page is a Jmol interactive view of the superposition of the three ALS models.
As above, but with the termini removed to provide a clearer view of the central region of the protein.
Jmol view of the superposition of the three models generated by Phyre2, HHpred and I-TASSER
References
1. Protein Folding Requires Crowd Control in a Simulated Cell. Jefferys BR, Kelley LA and Sternberg MJE Journal of Molecular Biology (2010) Volume 397, Issue 5, 16 April 2010, Pages 1329-1338.