Hello,
I have a query protein with 300 amino acids and I want to predict its homology models based on the template (saying templateA.pdb) with exactly same 300 amino acids (for testing purpose).
By using Robetta Server, what I did was listed below:
I pasted the query target sequence, and I selected comparative modelling option and set 1000 models.
I then uploaded the template PDB structure (templateA.pdb).
After ~2 days, I can get the result. I compared the retured top 5 models from Robetta Server and noticed that they have similar strucutures with little difference, which should be expected (because I used the simple templateA.pdb with the exactly same amino acids with qeury fasta sequence).
Now, I want to do the sample procedure by using RosettaCM protocol (following this tutorial: https://www.rosettacommons.org/docs/latest/application_documentation/structure_prediction/RosettaCM) where I used the same templateA.pdb to generate the threaded PDB structure (the threaded structure is totally same to templateA.pdb because the alignment between query sequence and the template sequence is 100% identical).
/Rosetta/3.12/bin/rosetta_scripts.static.linuxgccrelease @flags_hpc ...
The content of flags_hpc is listed below: Note I used a parallel computation from a external program like R calling (my computer has 16 cores) to generate 1000 different models (each of models has a different random number controled by -run:constant_seed and -run:jran. That is the reason I gave -nstruct 1 in the flags_hpc file instead of -nstruct 1000.
# i/o
-in:file:fasta WT.fasta
-nstruct 1
-parser:protocol rosetta_cm_hpc.xml
# relax options
-relax:dualspace
-relax:jump_move true
-default_max_cycles 200
-beta_cart
-hybridize:stage1_probability 1.0
The content of rosetta_cm_hpc.xml is listed below:
<ROSETTASCRIPTS>
<TASKOPERATIONS>
</TASKOPERATIONS>
<SCOREFXNS>
<ScoreFunction name="stage1" weights="score3" symmetric="0">
<Reweight scoretype="atom_pair_constraint" weight="0.1"/>
</ScoreFunction>
<ScoreFunction name="stage2" weights="score4_smooth_cart" symmetric="0">
<Reweight scoretype="atom_pair_constraint" weight="0.1"/>
</ScoreFunction>
<ScoreFunction name="fullatom" weights="beta_cart" symmetric="0">
<Reweight scoretype="atom_pair_constraint" weight="0.1"/>
</ScoreFunction>
</SCOREFXNS>
<FILTERS>
</FILTERS>
<MOVERS>
<Hybridize name="hybridize" stage1_scorefxn="stage1" stage2_scorefxn="stage2" fa_scorefxn="fullatom" batch="1" stage1_increase_cycles="1.0" stage2_increase_cycles="1.0">
<Template pdb="t2809_thread.pdb" cst_file="AUTO" weight="1.0" />
</Hybridize>
</MOVERS>
<APPLY_TO_POSE>
</APPLY_TO_POSE>
<PROTOCOLS>
<Add mover="hybridize"/>
</PROTOCOLS>
</ROSETTASCRIPTS>
Noted that in my RosettaCM protocal I did not provide the fragment files (such as 3-mer or 9-mer). Is that necessary to provide them as input of RosettaCM?
Finally, I used the energy_based_clustering.static.linuxgccrelease to cluster the generated 1000 models and extracted the first one from a cluster with maximum number of components.
I compared the one from my RosettaCM protocol with the one from Robetta Server, but unfortunately the two structrues are more different, more than what I expected.
I suppose my built model should be similar to the one from Robetta Server, but it is not.
Can you give me some suggestions to get the similar structures between the RosettaCM protocol and Robetta Server by giving the same inputs for both of them.
Or can you help check the options for RosettaCM protocol I used and help me figure out how can I setup them?
Thanks.