How is the helix axis calculated in 3DNA

[wwang02]

How does 3DNA define a helix? If I want to calculate a helix axis of a DNA, what position should I start to extend the local vector? and how should I calculate the base coordinates, by averaging all the atom coordinates of base pairs?

I need to know this, because I want to implement the analysis method of 3DNA to my simulations which are not done by the all-atom fashion. Does anyone know whether there exists any publications where I can refer to?

Thank you

[xiangjun]

In 3DNA, the global linear helical axis is defined as in SCHNAaP/SCHNArP, which follows NewHelix. Basically, equivalent C1' and RN9/YN1 atom pairs along each strand are employed in a least-squares fitting procedure to derive the helix. See "Structure and conformation of helical nucleic acids: analysis program (SCHNAaP)" (http://www.ncbi.nlm.nih.gov/pubmed/9356255) for details. The source code of SCHNAaP/SCHNArP is available --- see the Section "SCHNAaP/SCHNArP" (http://forum.x3dna.org/schnaap-schnarp/).

HTH,

Xiang-Jun
 

[wwang02]

If the DNA simulation is done in a coarse grain fashion, can I calculate the center of mass of triads and build up local unit vectors, and average the all the unit vectors to define the helix axis? Do you think that would be close enough to the method you use in the paper?

[xiangjun]

It is hard to say anything specific. Why not you have a try and see what happens?

Note that in addition to the 'global linear helical axis', 3DNA also defines a local helical axis for each dinucleotide step. See the 2003 3DNA NAR paper and the doc "tech_details.pdf" distributed with 3DNA v2.1 for details. You may find them useful.

Xiang-Jun

[wwang02]

Thank you for replies!