Netiquette · Download · News · Gallery · Homepage · DSSR Manual · G-quadruplexes · DSSR-Jmol · DSSR-PyMOL · DSSR Licensing · Video Overview· RNA Covers

Author Topic: Wrong junctions  (Read 36733 times)

Offline lijun

  • non-commercial
  • with-posts
  • *
  • Posts: 11
    • View Profile
Wrong junctions
« on: September 17, 2019, 11:48:56 am »
Hi Xiang-jun,
     When I used  dssr to extract junctions, I found the junctions for some RNAs share the same stem.
     For example, for RNA 4wsm, the output is:
          1* 3-way junction: nts=10; [0,4,0]; linked by [#100,#101,#101]
          60* 4-way junction: nts=17; [3,1,1,4]; linked by [#211,#212,#282,#212]
          76* 5-way junction: nts=14; [0,0,1,3,0]; linked by [#332,#333,#335,#333,#334]

      I used the latest version (1.9.6) of dssr, and the command as follows:
          x3dna-dssr -i=4wsm.cif -auxfile=no --isolated-pair=not-in-loop -idstr=long -o=4wsm.out
 
        Thanks and best wishes,
Jun
     

     

Offline xiangjun

  • Administrator
  • with-posts
  • *****
  • Posts: 1650
    • View Profile
    • 3DNA homepage
Re: Wrong junctions
« Reply #1 on: September 17, 2019, 01:56:52 pm »
Hi Jun,

A junction loop derived by DSSR may share the same stem when pseudoknots are involved, as emphased in the DSSR paper (https://doi.org/10.1093/nar/gkv716). See for example, Figure 4 for the env22 twister ribozyme, PDB id: 4rge. Generally speaking, this is a unique feature, instead of a bug, of DSSR. Such junctions are noted with a suffix * after the serial number. One can get rid of such loops by specifying the --nested option.

I've a quick look of your reported case on 4wsm, and noticed that there may be a bug in DSSR. The first case (a three way junction) should not be there. The issue is due to stem#101 (with 2 base pairs) which has a highly distorted geometry, and judged as parallel by DSSR. I will look into the issue further, and get back to you soon.

In the meantime, you could take it as a special case, or simply remove junctions with a * suffix.

Thanks for reporting the issue!

Xiang-Jun
« Last Edit: September 17, 2019, 01:58:41 pm by xiangjun »

Offline lijun

  • non-commercial
  • with-posts
  • *
  • Posts: 11
    • View Profile
Re: Wrong junctions
« Reply #2 on: September 17, 2019, 04:05:43 pm »
Hi Xiang-jun,
      Even with --nested option, some RNAs still share the same helix. They are 3j7r, 5mrc, 5mre, 5mrf, 6ost.
      Best,
Jun

Offline xiangjun

  • Administrator
  • with-posts
  • *****
  • Posts: 1650
    • View Profile
    • 3DNA homepage
Re: Wrong junctions
« Reply #3 on: September 17, 2019, 04:49:28 pm »
Hi Jun

Please provide more details with the additional PDB entries, as you did for 4wsm. I believe the underlying issue is similar, most likely associated with extreme cases with distorted structures. The more detailed cases you provide, the better I can test/validate the fixes for the next DSSR release.

Best regards,

Xiang-Jun

Offline xiangjun

  • Administrator
  • with-posts
  • *****
  • Posts: 1650
    • View Profile
    • 3DNA homepage
Re: Wrong junctions
« Reply #4 on: September 17, 2019, 11:44:51 pm »
Hi Jun,

I've updated DSSR (still labelled v1.9.6-2019sep16). It should have fixed the junction issue (in the first 3-way junction case) you observed in PDB entry 4wsm. See also my note on junctions with pseudoknots.

Please have a try and report back how it goes. In reporting any further issues, please remain to be specific.

Best regards,

Xiang-Jun

Offline xiangjun

  • Administrator
  • with-posts
  • *****
  • Posts: 1650
    • View Profile
    • 3DNA homepage
Re: Wrong junctions
« Reply #5 on: September 30, 2019, 10:39:52 pm »
As a follow-up, I've released DSSR v1.9.7-2019oct01 that's fixed the reported issue.

Xiang-Jun

 

Funded by the NIH R24GM153869 grant on X3DNA-DSSR, an NIGMS National Resource for Structural Bioinformatics of Nucleic Acids

Created and maintained by Dr. Xiang-Jun Lu, Department of Biological Sciences, Columbia University