plotting the helical axis along curved helices

[Di_Liu]

Hi Xiang-Jun,

I tried to extract the coordinates of the origin points of the base-pair reference frame from the json file. It appears working for DNA structures; but for RNA, the points form a spiral around the helical axis (see attached image), as we would expect due to the differences of how the axis passes through the base-pairs in B- and A-form helices.

Thanks,

Di

[xiangjun]

Hi Di,

Thanks for posting on the Forum. Indeed the origins of base-pairs are centered within a pair, as defined in the standard base reference frame. For B-form DNA where the helical axis passes through and is perpendicular to base pairs, the line connecting bp origins appears as expected. For A-form DNA or RNA (which is in A-form), the helical axis passes through the central hole where the bps (and their origins) spiral around. See "Figure 4. Influence of non‐zero Slide and Roll at sequential dimer steps on overall DNA helical conformation" of the 2003 3DNA paper in NAR.

DSSR does output a linear helical axis when a helical segment is not too strongly curved. See the 2015 DSSR paper "Figure 2 -- analysis of the yeast phenylalanine tRNA (1ehz)" for an example. You could also run the following commands:

Code: [Select]

x3dna-dssr -i=1ehz.pdb --helical-axis
pymol 1ehz.pdb dssr-helicalAxes.pdb
# within PyMOL: as lines; png 1ehz-helices.png
You will see an image as attached below.

However, DSSR currently does not fit a smooth curvilinear helical axis around an arbitrary shape, e.g., a DNA circle. In principle, DSSR can fit a mini-helical axis for each base-pair step (i.e., a 2-bp segment) and then perform a b-spline interpolation. I'm open to suggestions and welcome collaborations to pursue this topic further.

Best regards,

Xiang-Jun

[Di_Liu]

Thanks, Xiang-Jun!

Do you have any idea of how to easily find the axis of each 2-bp segment of a helix?

Also, I think an easier solution for A-form helix is to do a shift of the origin in the plane of the reference frame so that the shifted origin is where the axis passes through the plane.

Di

[xiangjun]

Hi Di,

Do you have any idea of how to easily find the axis of each 2-bp segment of a helix?

The info is within DSSR, but not exposed. I'm considering to add this feature in DSSR JSON output for easy parsing. For WC-like pairs, things are not that complicated. However, with there are subtitles with non-Watson-Crick pairs, e.g., Hoogsteen and reverse Hoogsteen base pairs.

Also, I think an easier solution for A-form helix is to do a shift of the origin in the plane of the reference frame so that the shifted origin is where the axis passes through the plane.

See "Worked examples on base-pair parameters" in the DSSR Pro User Manual. especially Session "6 Local helical parameters". The vectors o1_h and o2_h are what you need. They are not simple shifts of the origin in the reference frame.

This thread actually prompt me to refine the detailed algorithmic descritpion and get the content published. They are the real meat of 3DNA!

Best regards,

Xiang-Jun