I am trying to use FloppyTail to determine a C-terminus tail structure of 6 consecutive His residues. However, several residues before the tail also experience conformation change though the majority is remained the same. Can I ask is there a way to only optimise the tail structure?
I am not sure if the constraint file can solve this. As I understand, the constraint file is more of a modification to the relax algorithm.
Thank you very much.
The following is my options file
FloppyTail is not meant for the question of, "I have a terminal tail that I think is well-structured, what does that structure look like?" FloppyTail IS meant to answer, "I have a flexible region that I expect to be fairly unstructured, and would like to widely sample possible conformations, to interpret in tandem with other experimental data".
I don't know if 6xHis tags are generally unstructured. FloppyTail will be fine for sampling possible conformations of your protein tail plus the 6xHis tag.
FloppyTail does not add residues - you need to start with a model with all residues present (try the loop model demo I discussed in response to your other question).
As far as constraints go - constraints are just arbitrary, user-defined modifications to the scorefunction. You, the experimenter, might know that residues X and Y interact, so you can constrain them to be nearby. If you have reason to believe the tail's real conformation has a particular feature, then constraints are a great way to inform the scorefunction of that fact and bias your modeling to sample structures with the feature.
That's not a real options file. You need to correct the -database and -s paths. You'll want to correct flexible_start_resnum and flexible_chain to your actual structure. Also, you need *way* more cycles for perturb_cycles and refine_cycles - how many you need is empirical, but I'd start with something like 5000 and 10000. Same for nstruct, you surely want more than one model. I should point out that FloppyTail is not meant to be run on a desktop computer...
It occurs to me that 6xHis tags probably have a canonical structure (if they adopt a structure at all) that you can look up somehow.
Thank you very much for your instructions. I am sorry I did illustrate my question clearly enough.
1) Actually, I am trying to use 4KMT (PDB number) as a homology model. However, the original structure downloaded from website does not resolve the end of the heavy chain (HC) of 6 Histidines. Therefore, I think FloppyTail is a good method to estimate the possible structure alternatives. I use the nearest His coordinates (i.e. residue 206) to fill the residues 223 to 228 before invoking FloppyTail. Can I ask if I have understood it correctly?
2) Thank you for the explanation about constraints. Now I realise the constraints are somewhat based on subjective interpretation for the interactions involved. In my case, I would say I have no idea about this 4KMT. So constraints might be not feasible at the moment. But is there some way that only explores the tail conformation without changing the rest?
3) For the options file, as you said, I will customise the -database and -s paths. The flexible_start_resnum and flexible_chain already correspond to the 4KMT.
4) In terms of the cycle number and -struct, the thing is, I am using 4KMT to estimate my own antibody. Actually, there is not any sequence similarity between 4KMT and my antibody with regard to residue 223-228. Therefore, I think there is not so much meaning to fully explore the possible conformation within this tail as I will always explore it after replacing to my own sequence (i.e. comparative modeling). Probably I should do a little bit more, e.g.
but I do not need to do that much. (Besides, I am still waiting our university to install Rosetta on the cluster, and I have to use my own laptop.) How do you think of this?
You can try FloppyTail to consider what a His tag is going to do, but since terminal 6xHis tags are so common, I think there may be literature that addresses it instead.
1) I don't think I've been quite clear. Copying his206 as 223-228 will give a valid starting structure for CCD-based loop modeling, but unfortunately not FloppyTail. FloppyTail is only smart enough to modify torsions, but not repair the bond lengths and angles. CCD modeling with the 'extended' option WILL fix the bond lengths/angles, which will produce an intermediate structure you can then feed in to FloppyTail.
4) I don't think cycle counts of 500 and 300 will be enough to really sample very much. FloppyTail is a very inefficient Monte Carlo search. I guess you can try it.
Cluster - you don't need your university to install Rosetta for you. You ought to be able to install it to your userspace on the cluster the same way you install it to userspace on your laptop. Certainly when I use Rosetta on a cluster, I just install it for myself.
Thank you very much.
So I will also use "loop_length_changing" for the tail 223-228. But I will not use the inefficient FloopyTail as the tail has no similarity to my own antibody.
I will ask our IT people if Rosetta can be installed to my own userspace.