Recent Posts

1

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by xiangjun on Today at 01:34:37 pm »

Hi Jiaolong,

Thank you for providing the technical details and verification of the DSSR implementation for the classification of L-sugars. Detailed responses like this help improve the software and set a standard for community interaction.

I compiled four examples based on the information you provided to illustrate how the signs of the base-pair parameters are defined. See the Practical Guide associated with the 2020 DSSR-PyMOL paper for details. The base blocks and the standard base-reference frame attached to each base clarify the sign convention of base-pair and step parameters. PyMOL session files are attached so users can easily verify the results.

DSSR is under active development, and community interactions improve the software. I address questions and bug reports quickly. Users are encouraged to seek clarification regarding execution errors or unsupported features on the forum for their own benefit and that of the community as a whole.

Best regards,

Xiang-Jun

2

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by JiaolongBao on July 06, 2026, 10:31:11 pm »

Hi Community,

The DSSR has been updated to v2.9.1 and it could be requested license at http://innovation.columbia.edu/technologies/CU20391. Regarding the update, the newest version is able to classify L-sugar to correct puckering type while retaining the original phase angle for L-sugar. The sugar class classification for L-sugar is omitted due to some compatibility reason with D-sugar.

Many thanks to Xiang-Jun's continued dedication to DSSR which makes it more and more powerful.

Jiaolong

3

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by JiaolongBao on June 17, 2026, 09:08:34 pm »

Hi Xiang-Jun,

Certainly! I will email you for a video meeting.

Many Thanks,
Jiaolong

4

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by xiangjun on June 17, 2026, 06:50:45 am »

Hi Jiaolong,

Thanks for the detailed explanation—this kind of feedback is exactly what I have been looking for.

I will certainly investigate the sign convention further. In this post, I would like to address the sugar-class issue. Currently, the L-sugar class is determined based on the pseudorotation phase angle, using the exact same criteria as for D-sugars. I have attached the dssr-torsions.txt output for 1BNA_L.pdb. Honestly, I had not paid close attention to this specific aspect of L-DNA before. Please let me know the correct way to determine the sugar class in these cases.

For this type of discussion, a phone call or a video meeting might be more convenient. Please let me know if you would like to connect online or over the line to discuss this topic further.

Best regards,

Xiang-Jun

5

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by JiaolongBao on June 17, 2026, 02:36:37 am »

Hi Xiang-Jun,

Thanks for your reply and continued effort looking in to source code of DSSR.

I appreciate that you provide the newest version DSSR output regarding 1BNA_L.pdb. In this case, the sign inversion pattern of base-pair parameter is the same as v2.6.0. However, the sign inversion pattern of step/helical bp parameters looks different with v2.6.0 in terms of shift, slide, tilt, roll, x-disp, y-disp, inclination and tip. Out of curiosity, I also downloaded SCHNAaP and X3DNA source code and use them to analyze the 1BNA_L.pdb, which gives me a more "reasonable" sign inversion pattern. I summarized the sign-inversion patterns in the three attached tables. Here, “Yes” means that the sign of the corresponding parameter is reversed between 1BNA.pdb and 1BNA_L.pdb, while “No” means that the sign is preserved. along with SCHNAaP, X3DNA v2.4, DSSR v2.6 and v2.9 for your reference.

As you can see in the tables, SCHNAaP and X3DNA share the same sign inversion pattern for base-pair and step/helical parameters which looks more intuitive to me. Interestingly, DSSR v2.9 now gives the same sign-inversion pattern as SCHNAaP and X3DNA for the step parameters and helical base-pair parameters. However, for the base-pair parameters, DSSR v2.9 still follows the same pattern as DSSR v2.6, which differs from SCHNAaP and X3DNA for shear, stagger, buckle, and opening.

Based on my understanding of SCHNAaP paper (which has very detailed explanation on how to calculate those parameters, although it is outdated since DSSR is using base reference frame), the calculation of base-pair and step/helical parameters is very similar and I believe mirror-image transformation will lead to same sign inversion pattern. I feel the sign inversion of parameters heavily depends on the axis direction and also which axis is determined by right-handed rule. Anyway, I'm more interested in DSSR how to handle this scenario.

Besides, I also notice when DSSR judge the "sugar-class" of L-sugar may have some issues. I guess DSSR is based on the phase angle to determine the sugar class of sugar which is totally fine to D-sugar while it may not be suitable for L-sugar. I believe the chirality inversion of sugar will not affect the sugar class but phase angle should be definetely inversed. For example, DSSR determine the sugar class of nucleotides in 1BNA_L.pdb is C3'-endo while it should be C2'-endo no matter in D or L form B-DNA.

Code: [Select]

Summary of structural features of 24 nucleotides
  Note: the first five columns are: (1) serial number, (2) one-letter
    shorthand name, (3) dbn, (4) id string, (5) rmsd (~zero) of base
    ring atoms fitted against those in a standard base reference
    frame. The sixth (last) column contains a comma-separated list of
    features whose meanings are mostly self-explanatory, except for:
      turn: angle C1'(i-1)--C1'(i)--C1'(i+1) < 90 degrees
      break: no backbone linkage between O3'(i-1) and P(i)
   1  C ( A.DC1    0.020  anti,~C3'-endo,canonical,non-pair-contact,helix-end,stem-end
   2  G ( A.DG2    0.025  anti,~C3'-endo,canonical,non-pair-contact,helix,stem
   3  C ( A.DC3    0.024  anti,canonical,non-pair-contact,helix,stem
   4  G ( A.DG4    0.023  anti,~C3'-endo,canonical,non-pair-contact,helix,stem
   5  A ( A.DA5    0.016  anti,~C3'-endo,canonical,non-pair-contact,helix,stem
   ................

I hope our discussion can make DSSR more capable for analysis of L-DNA/RNA. Thanks!

Best regards,
Jiaolong

6

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by xiangjun on June 16, 2026, 08:30:02 am »

Hi JiaolongBao,

Thank you for your follow-up on this topic. The analysis of L-DNA (with mirror-image sugar-phosphate backbones) is a new DSSR feature, added in response to user requests. DSSR can identify the mirror-image sugar components and classify the corresponding steps as L-DNA with a negative twist angle. Running DSSR on your 1BNA_L.pdb file yields the following results:

Code: [Select]

# x3dna-dssr -i=1BNA_L.pdb --more
  helix#1[1] bps=12
      strand-1 5'-CGCGAATTCGCG-3'
       bp-type    ||||||||||||
      strand-2 3'-GCGCTTAAGCGC-5'
      helix-form  LLLLLLLLLLL
    helical-rise:   3.35(0.44)
    helical-radius: 9.37(0.98)
    helical-axis:    0.088    -0.275    -0.958
       point-one:  -16.795    25.817    26.180
       point-two:  -13.557    15.707    -9.067
   1 A.DC1          B.DG24         C-G WC           19-XIX    cWW  cW-W
        bp1-pars:* [-0.42    -0.27    -0.06    -2.76    14.18    -3.67]
       step-pars:  [0.36     0.15     3.52     -3.40    -6.44    -40.31]
       heli-pars:  [0.55     0.11     3.52     9.26     -4.88    -40.94]
        bp2-pars:* [-0.02    -0.27    -0.25    4.45     10.83    -4.02]
       C1'-based:                rise=3.52                 twist=-38.57
       C1'-based:              h-rise=3.52               h-twist=-39.23
...

Regarding the sign inversion pattern of the base-pair and step/helical parameters, you are the first user to raise this question. In the current implementation, certain base-pair parameters are reversed due to the adjustment of the base reference frame, and the corresponding step/helical parameters follow those adjustments. I will examine the source code details and follow up on this topic soon.

From a user perspective, what behavior do you consider most logical? As the DSSR developer, I want the tool to meet user expectations while remaining mathematically consistent. The current implementation satisfies the requirements of Di and Gengshi, who initially requested this feature.

Best regards,

Xiang-Jun

7

RNA structures (DSSR) / Re: Can 3DNA DSSR handle Left-handed DNA?

« Last post by JiaolongBao on June 15, 2026, 09:58:56 pm »

Dear Xiang-Jun,

I am currently using DSSR to help me figure out the base pair, step base pair, helical base pair parameters difference between D and L form B-DNA. I just notice this post is regarding L-form DNA recognition and realize you update DSSR in v2.7.1 to acommondate L-DNA is a sense while I am still using a bit outdated version DSSR v2.6.0-2025jul24 to analyze a native B-DNA duplex and its mirror-image counterpart.

Specifically, I used PDB entry 1BNA as the native structure and generated a mirror-image structure by reflecting the X coordinates(x -> -x). I attached the JSON format output file in this post, but anyway I list the sign inversion conclusion here:

Code: [Select]

Shear      : no sign inversion
Stretch    : no sign inversion
Stagger    : sign inversion
Buckle     : sign inversion
Propeller  : sign inversion
Opening    : no sign inversion

Code: [Select]

Shift  : no sign inversion
Slide  : sign inversion
Rise   : no sign inversion
Tilt   : sign inversion
Roll   : no sign inversion
Twist  : sign inversion

Code: [Select]

x-displacement : no sign inversion
y-displacement : sign inversion
helical rise   : no sign inversion
inclination    : sign inversion
tip            : no sign inversion
helical twist  : sign inversionHere which makes me feel confused is that the sign inversion pattern of the base pair parameters is different from that of the step and helical parameters. I understand that if the inversion pattern of bp parameters apply to step/helical bp parameter, such as Twist and Helical Twist will be "no sign inversion" which is impossible since the most distinct difference between native and mirror-image double helix is right and left handed. I looked into X3DNA and also SCHNAaP paper which has detailed description about how to calculate the parameters which makes believe the sign inversion pattern should be same . Probably DSSR has some certain axis direction inversion when calculate step/helical bp parameters which I overlooked.

I am also wondering the DSSR V2.7.1 or the newest version will give out the same conclusion or there is some certain adaption when encounter L-form when calculating the parameters. I will be very grateful to your help and sorry if I misunderstood something.

8

RNA structures (DSSR) / Re: overlapping area calculation - stacking interactions

« Last post by Agnieszka on June 08, 2026, 01:25:39 pm »

Dear Xiang-Jun,
Thank you very much! I appreciate your help. I have already check how it looks for my reference structure and it is perfect.
Best regards,
Aga

9

RNA structures (DSSR) / Re: overlapping area calculation - stacking interactions

« Last post by xiangjun on June 08, 2026, 10:09:05 am »

Hi Agnieszka,

wDSSR has been updated with overlapping area calculation for 3DNA output. See 355d for an example.

Please let me know if you have any questions.

Best regards,

Xiang-Jun

10

RNA structures (DSSR) / Re: overlapping area calculation - stacking interactions

« Last post by Agnieszka on June 03, 2026, 12:42:34 pm »

Yes, please. That would be convenient to have it as downloadable wDSSR file.
Thank you,
Agnieszka