Hello, everyone!
I'm aiming to study the impact of two AA substitutions on 3D conformation in a 29AA-peptide for which there is no available 3D structure in PDB.
I came accross with a previous report investigating a similar issue which employed an homology-based approach (Phyre2) to obtain 3D models, but then I wondered if this approach couldn't bias the analysis by minimize the effect of variation, and if de novo models could work better for this purpose by considering force field changes caused by AA substitution. Does this hipothesis makes sense? If it does, could anyone indicate me a reference supporting this.
Additional question: I've already obtained models for the wild-type structure using the homology-based approache (Phyre2) and knowledge-based de novo approaches (QUARK, I-TASSER and Rosetta). And also obtained refined models using FG-MD, ModRefiner and PREFMD and analyzed raw and refined structures using PROCHEK and PROSA. Which parameters should I consider more when choosing the more reliable approach (that which generate more reliable structures)?
Hi Glauco,
since you have already prepared a model of the 29 aa peptide. Try to pick the model generated from various softwares, depending on its RAMACHANDRAN plot.
As you have mentioned, you have performed PROCHEK analysis on all models so might be able to pick the right model.
After selecting the model go for MD studies using GROMACS. MD simulation of 50-100 ns are sufficient enough to minimize the conformation of peptides (wild & mutated).
After MD simulation, you can overlap the two models and calculated rmsd of the models. Also structural changes in the mutated residues can be analyzed.
Hi Glauco - In my opinion better try to find a collaborator who can perform the conformational assignment using NMR - much less speculative than the modelling and it will help you to attract more relevant journals.
Tanuj Sharma, thank you for your answer!
Actually, RAMACHANDRAN plots returned very similar results for the best models obtained from different softwares. Is there any other critical parameter to pick a model and use the same software for future prediction of mutated structures?
Philippe Derreumaux, thank you for your answer.
Actually I also tested PEP-FOLD, just forgot to include it in my question. Unfortunatelly, it returned the more bizarre structure compared with other softwares and based on the biological function of the peptide.
Hi Glauco,
you can use SAVES server to perform various type of computational analysis on the model. Also whatIF server can be used to predict behavior of mutated peptides.
regards
Tanuj
Thank you, Tanuj Sharma!
Philippe Derreumaux, what actually happened when I used PEP-FOLD 3.1 was that it returned some structures with two alpha helix, while all other softwares returned mainly structures composed of a single alpha helix, which is consistent was consistent with secondary structure predictors and its' biological function (a signal peptide). But now reanalyzing data obtained by PEP-FOLD, I have two models with this archtecture, and when reanalyzing them using PROCHECK, despite having Ramachandran plots with less aminoacids in "most favoured regions" (90 and 80%, while best models from other softwares had 95%), they had better G-factors for dihedral angles and main-chain covalent forces, showing no or little deviation from -0.5. May these models be more realistic than those with better Ramachandran plots but with worst G-facotrs?
Sorry, but is your email address correct? There isn't something before @ibpc.fr?
Thank you so much for atention!
Due to their small size, peptides are faster to analyze by molecular dynamics followed by DSSP to summarize the frequency of secondary structures per residue. However, this will need a lot of resources if you have a great number of mutants.
I've just received predicted structures from Robetta server (it took 1 week to complete the job). The two best models showed 95% of aminoacids in the most favored regions in Ramachandran plot and no unusual scores for dihedral angles or main chain covalent forces.
- Badges
- Science topic
- Similar topics
- Chemistry
- Biochemistry