Building a bridge between Curves+ and 3DNA

In the field of nucleic acid structural analysis, it seems fair to say that Curves+ and 3DNA are nowadays the top two choices. To the best of my knowledge, these two programs are also the only ones that confirm to the standard base reference frame. Moreover, as noted in my previous post, Curves+ and 3DNA are “constructive competitors” with complementary functionality: Curves+ is unique in providing a curvilinear helical axis, a bending analysis, a full description of groove widths and depths and its seamless integration to the analysis of molecular dynamics trajectories, while 3DNA’s strength lies in its cohesrent approach combining analysis, rebuilding, and visualization into one package.

Given the complementarity between Curve+ and 3DNA, it makes sense to build a ‘bridge’ between the two so users can easily take advantage of both programs. Starting from 3DNA v1.5, find_pair has the -c option to generate input for Curves directly from a PDB file. Over the years, this option appears to have received little attention — at least, I am not aware of any literature reference to it. Now, the updated Curves+ program has introduced the new lib name list variable, among other changes. I have thus added the -curves+ option (abbreviation -c+) to find_pair to make its output compatible with Curves+.

As always, the point/process is best illustrated with an example — here with the Dickerson B-DNA dodecamer solved at high resolution by Williams et al. (PDB entry 355d).

find_pair -c+ 355d.pdb 355d-curves+.inp

The generated file 355d-curves+.inp has the following content:

&inp file=355d.pdb,
    2    1   -1    0    0
    1    2    3    4    5    6    7    8    9   10   11   12
   24   23   22   21   20   19   18   17   16   15   14   13

which can be fed into Curves+ as below,

Cur+ < 355d-curves+.inp

The four output files are: 355d.cda, 355d.lis, 355d_X.pdb and 355d_b.pdb.

Please note the followings:

  • The environment variable CURVES_PLUS_STDLIB should be set, pointing to the directory where Curves+ is installed (containing files standard_b.lib and standard_s.lib). In the example above, CURVES_PLUS_STDLIB is set to /Users/xiangjun/Curves+.
  • The find_pair -c+ option (currently) is applicable only to double helical DNA/RNA structures, the most common application scenario.
  • The -c+ option ignores HETATM records, in accordance with Curves+ where proteins, water and HETATM are automatically removed at input. (see Curves+ user manual, section Input data)
  • To run Cur+ < 355d-curves+.inp again in the same folder, the four output files must first be deleted (e.g., rm -f 355d.cda 355d.lis 355d_[Xb].pdb). This is best taken care of via a script.

Obviously, the nucleic acid structure community benefits the most to have both Curves+ and 3DNA at its disposal and be able to easily switch between them — hopefully, the find_pair -c+ option would serve as such a ‘bridge’.



Dear Author,

I followed this webpage instruction but when I input “Cur+ < 355d-curves+.inp”, I got an error message: “Cur+: Command not found”, though I install Curves+ successfully, I even see a Cur+ icon in the folder.

Do you have any clue what was wrong?


— Xiao Lei · 2017-05-14 12:24 · #


Thanks for stopping by and leaving a comment. Since you have the error message “Cur+: Command not found”, you need to make sure that Curves+ has been installed properly. It is a independent program from 3DNA. You should run the examples distributed with Curves+ to verrify your installation before making sense of the 3DNA-Curves+ bridge enabled by the “find_pair -c+” option.

— Xiangjun · 2017-05-14 12:32 · #



Thank you for printing this article from Please do not forget to visit back for more 3DNA-related information. — Xiang-Jun Lu