Change in Gibbs free energy (delta delta G) of RNA

[Anonymous]

Hello,

I’m trying to load a .pdb file for an RNA molecule, change a single base in the molecule, and calculate the change in free energy. I was looking at this tutorial (http://graylab.jhu.edu/pyrosetta/downloads/documentation/PyRosetta_Tutorial_2012.pdf) and it doesn’t work with newer versions of PyRosetta. Has anyone else done something like this? The documentation hasn’t been very helpful so far for RNA (though it’s relatively extensive for protein).

 

[Anonymous]

For updated tutorial material that should work for PyRosetta4, see http://www.pyrosetta.org/tutorials  

The documentation about RNA is rather spotty because investigating RNA with Rosetta is rather new, with comparatively fewer labs working on it, when compared with the work on proteins. To wit, when I asked someone in the Das lab if they had any recommendations about computing ddGs of mutation in RNA, they had this to say:

In brief, ‘no’. In long form: no, because it’s either trivial or a phenomenally challenging calculation.

When mutating a base in a helix, you pretty much can use the difference in secondary structure energy.

When mutating a base in a loop, you have a really hard problem that the [as yet unpublished] ‘thermal sampler’ app is intended to begin to approach solving.

That’s not to say that you can’t attempt it, just realize that you aren’t going to get all that rigourous results out of the protocol. The basic approach of working with RNA should be more-or-less the same as working with proteins, with the big caveat that most of the protocols haven’t been benchmarked with RNA, so a) you might run into edge cases where a protocol assumes that the structure is protein, and fails because of it or b) the protocol might be able to proceed formalistically without errors, but because of the structural differences in protein and RNA a protocol optimized for protein might not give decent results for RNA.

If you do put together a protocol using RNA, I *highly* recommend benchmarking it first on some “positive controls” – or systems where you know the answer you’re attempting to find for your system of interest. (The closer these systems are to your unknown system, the better.) This will give you a sense if the protocol is working decently, or is just spitting back garbage numbers at you.

 

[Anonymous]

For updated tutorial material that should work for PyRosetta4, see http://www.pyrosetta.org/tutorials  

The documentation about RNA is rather spotty because investigating RNA with Rosetta is rather new, with comparatively fewer labs working on it, when compared with the work on proteins. To wit, when I asked someone in the Das lab if they had any recommendations about computing ddGs of mutation in RNA, they had this to say:

In brief, ‘no’. In long form: no, because it’s either trivial or a phenomenally challenging calculation.

When mutating a base in a helix, you pretty much can use the difference in secondary structure energy.

When mutating a base in a loop, you have a really hard problem that the [as yet unpublished] ‘thermal sampler’ app is intended to begin to approach solving.

That’s not to say that you can’t attempt it, just realize that you aren’t going to get all that rigourous results out of the protocol. The basic approach of working with RNA should be more-or-less the same as working with proteins, with the big caveat that most of the protocols haven’t been benchmarked with RNA, so a) you might run into edge cases where a protocol assumes that the structure is protein, and fails because of it or b) the protocol might be able to proceed formalistically without errors, but because of the structural differences in protein and RNA a protocol optimized for protein might not give decent results for RNA.

If you do put together a protocol using RNA, I *highly* recommend benchmarking it first on some “positive controls” – or systems where you know the answer you’re attempting to find for your system of interest. (The closer these systems are to your unknown system, the better.) This will give you a sense if the protocol is working decently, or is just spitting back garbage numbers at you.

 

[Anonymous]

For updated tutorial material that should work for PyRosetta4, see http://www.pyrosetta.org/tutorials  

The documentation about RNA is rather spotty because investigating RNA with Rosetta is rather new, with comparatively fewer labs working on it, when compared with the work on proteins. To wit, when I asked someone in the Das lab if they had any recommendations about computing ddGs of mutation in RNA, they had this to say:

In brief, ‘no’. In long form: no, because it’s either trivial or a phenomenally challenging calculation.

When mutating a base in a helix, you pretty much can use the difference in secondary structure energy.

When mutating a base in a loop, you have a really hard problem that the [as yet unpublished] ‘thermal sampler’ app is intended to begin to approach solving.

That’s not to say that you can’t attempt it, just realize that you aren’t going to get all that rigourous results out of the protocol. The basic approach of working with RNA should be more-or-less the same as working with proteins, with the big caveat that most of the protocols haven’t been benchmarked with RNA, so a) you might run into edge cases where a protocol assumes that the structure is protein, and fails because of it or b) the protocol might be able to proceed formalistically without errors, but because of the structural differences in protein and RNA a protocol optimized for protein might not give decent results for RNA.

If you do put together a protocol using RNA, I *highly* recommend benchmarking it first on some “positive controls” – or systems where you know the answer you’re attempting to find for your system of interest. (The closer these systems are to your unknown system, the better.) This will give you a sense if the protocol is working decently, or is just spitting back garbage numbers at you.

 

[Author]

Posts