Single Strand DNA (ssDNA) 3D structure prediction

[mkoohim]

I have an ssDNA and want to predict its 3D structure. Can I use 3DNA?
The secondary structure of my sequence is attached. For RNA we have some software such as Rosetta to predict 3D structure but unfortunately I could not find any tools that predict tertiary structure of ssDNA? Please guide me, if 3DNA can predict this structure. If 3DNA can not do it, is there any other software to do it?   

Thanks alot

[xiangjun]

See the thread "changing secondary structure to tertiary structure of ssDNA".

For RNA we have some software such as Rosetta to predict 3D structure

Does it make sense to start from the same secondary structure you have, but simply change Ts to Us, and predict a 3D RNA structure using Rosetta etc. Then you can delete O2' atoms, and mutate Us to Ts with mutate_bases that will preserve backbone conformation and base-pairing geometry.

Xiang-Jun

[mkoohim]

Dear Xiangjun,
With this displacement I can use Rosetta etc. But I have thought the final result that make by this method is not very good. Because as you know, the software calculate the energy between nucleic acid and try to  minimize the whole structure of energy. In your opinion it is not important that  Ts are changed to Us and calculate the whole energy of structure with Us? For example seq1 and seq2 have a same 3D structure if we suppose seq1 is an ssDNA?

seq1: ATACCAATCTTAACC
seq2: AUACCAAUCUUAACC

Thank you very much,
Mohamad

[xiangjun]

No, I'm not saying or intending to imply that ssDNA and ssRNA are energetically equivalent. However, from a practical prospective, first it'd be better to have something than nothing. Second, a prediction is just a prediction, it is no guarantee that the "predicted" structure is "real" or even "meaningful": just think about how many predictions you can have from the same program with different settings, not to mention the different software tools available. Third, the recent article "RNA-Puzzles: A CASP-like evaluation of RNA three-dimensional structure prediction" [April 2012 [18 (4)] issue of the RNA journal] serves as a proof of my point. In theory (and in practice), a software can easily "predict" many more protein/DNA/RNA structures than the PDB has accumulated over the past 40 years.

Aside from all the augments above, 3DNA does not deal with "energetics" at all (in its current version, at least) -- it is purely geometrical.

Xiang-Jun

[mkoohim]

Dear Xiangjun,
Thank you very much for your advices. Also thank you for the excellent paper.

Best,

[priyamvada161]

hi. i have a similar query. i want to know the tertiary structure of a ssDNA aptamer. xiangjun could you please elaborate why you suggest deleting the O2 atoms. Also mkoohim were you able to find any other suitable software to predict the structure.

thanks
priyamvada

[xiangjun]

xiangjun could you please elaborate why you suggest deleting the O2 atoms.

Assuming one knows how how to predict a ssRNA structure with Rosetta etc, to convert to a corresponding ssDNA structure, I was suggesting to mutate U to T, and to remove O2' atoms (which are RNA specific). Does that make sense? Of course, a software/server for direct ssDNA structure prediction would be better.

Xiang-Jun

[mkoohim]

I followed xiangjun advice. I think it's the best way.