3DNA changes base-pair order

[armanf]

Dear All,
I am trying to use 3DNA to extract step parameters of a circular DNA from a MD simulation of 180 base-pairs.
I use find-pair and analyze commands to get cf_7methods file. Here I have attached 2 output files and the corresponding analysed files which only have 1 picoseconds in between but some problems occur:
1- in one of them number of recognized bse-pairs are less thatn the other.
2- For some base-pairs, the base-pair step parameters cannot be obtained and 3DNA gives "----". I tried to play with all parameters of the misc_3DNA.par but still there are some unknown parameters for some base-pairs.
It seems that for highly deformed bases or base-pairs 3DNA do not consider them as bases (base-pairs). Is there any way to ignore it?
3- Last but not the least, 3DNA changes order of base-pairs in the file. For example compare the sequence of base-pairs 160 to 180 in the files 29999.par and 30000.par . (the 30000.par is the correct sequence). Is this only a mistake in the written sequence or everything has been changed?

Thanks in advance
With the best regards
Arman

[xiangjun]

Hi Arman,

Thanks for using 3DNA and for posting your questions on the 3DNA Forum.

The issues you experienced when applying 3DNA for the analysis of molecular dynamics trajectories are well known, and there are a few things you can to solve them.

• How to fix missing (superfluous) base pairs identified by find_pair?
• How to properly use x3dna_ensemble?
• Read the 3DNA Jove paper which has a recipe on analyzing MD data
• Check do_x3dna

Basically, you need to prepare a template input file for each frame of the trajectory. The initial input structure can be generated using find_pair and manually edited as necessary.

Regarding the missing base-pair step parameters "----", see the analyze commandline help (note the -c option):

===========================================================================
NAME
        analyze - calculate nucleic acid structural parameters
SYNOPSIS
        analyze [OPTION] FILE ...
DESCRIPTION
        calculate various nucleic acid structural parameters (propeller,
        slide, roll, twist, backbone torsions etc.) from FILEs, and
        generate input to other utility programs. As of v2.1 with the
        -torsion option, it also provides detailed pseudo-torsions, Zp
        for single-stranded DNA/RNA structures, and classification of
        backbone BI/BII and base syn/anti-conformation
        -c      output structural parameters between helical regions
                ("----" by default). The same effect can be achieved by
                directly modifying the input file (change "9" or "1" to
                "0" in the third column of each base pair list.)
        -t      detailed (pseudo) torsions, BI/BII, syn/anti- and Zp
        -h      this help message (any non-recognized options will do)
INPUT
        given a PDB file "sample.pdb", the input to analyze can be most
        conveniently generated with the utility program find_pair:
        find_pair sample.pdb sample.inp
        an explicit input file (including 'stdin') must be specified.
EXAMPLES
        analyze sample.inp
        analyze sample1.inp sample2.inp sample3.inp
        find_pair sample.pdb stdout | analyze stdin
        find_pair sample.pdb stdout | analyze -c stdin
        analyze -t=6tna.tor 6tna.pdb
OUTPUT
        sample.out, auxiliary.par, bp_step.par, bp_helical.par,
        cf_7methods.par, ref_frames.dat, poc_haxis.r3d, stacking.pdb
        hstacking.pdb
SEE ALSO
        find_pair, rebuild, frame_mol, ex_str, stack2img, cehs
AUTHOR
        3DNA v2.1-2014dec22, created and maintained by Xiang-Jun Lu (PhD)

Please post questions/comments on the 3DNA Forum: http://forum.x3dna.org/
===========================================================================

HTH,

Xiang-Jun

[armanf]

Dear Xiang,
Thanks very much indeed for your comprehensive and prompt reply.
The command analyze -c was very helpful and one of my problems just solved.
Also thank you for sending me the links. I am already familiar with MD analysis of 3DNA and I just posted the two snapshots to clear my point.

But still my biggest concern remains that the order of base-pairs or maybe name of them is changing in different snapshots. For example base-pair 169 in '29999.par' is C-G but in '30000.par' is G-C. Do you know why that happens? If they were in a goos order, at least I could find out which base-pair is highly deformed that cannot be understood by 3DNA.

Thank you very much again
Arman

[xiangjun]

But still my biggest concern remains that the order of base-pairs or maybe name of them is changing in different snapshots. For example base-pair 169 in '29999.par' is C-G but in '30000.par' is G-C. Do you know why that happens? If they were in a goos order, at least I could find out which base-pair is highly deformed that cannot be understood by 3DNA.

The different order is because of the fact that the structures are distorted in different ways, and the 'find_pair' algorithm is not smart enough to figure out which way to go, and 'randomly' picks one based on its default settings. If you do your MD analysis as suggested, this won't become an issue: you need to generate an initial input using any frame with find_pair and edit it as necessary. Think of the scenario where you do not have access to find_pair.

On the other hand, you can make use of find_pair in an unconventional ways: the regions where the order changes should be where deformation occurs.

HTH,

Xiang-Jun

[armanf]

Dear Xiang,
Thank you so much.
The problem is that when I use a sane input file, generated by find-pair and use it as a template for x3dna_ensemble, at a snapshot I get this error and process stops:
Error in running command: '/mnt/store2/home/.../bin/analyze -c  temp_model.inp 2> msgfile'
Rerun the command w/o redirection for details

this is the command that I used:
x3dna_ensemble analyze -b inp.inp -l pdb_list.txt -o ensemble.out

and I have changed values in the misc_3DNA.par file to avoid former unknown base-pairs problems.
Can you please also help me with that?
I really appreciate it.
Best

[xiangjun]

Please provide details so the error message can be reproduced.

Thanks,

Xiang-Jun

[armanf]

Thanks.
Here I have attached three pdb files that I am using them for analysis. the first and last one can be analyzed without any problems. But if you add them in a pdb_list file, when it starts to analyze frame1067.pdb the error message pops out. I also attached the inp.inp file which is the template for base-pairs generated by find_pair command.
I found that when I use find_pair for frame1067.pdb even with the modified variables of misc_3DNA.par, still it does not recognize one of the base-pairs so that could be the source of error. But I do not know what else could I do to resolve it.

Regards

[xiangjun]

Thanks for attaching the files.

You've touched on a subtle point in structure frame1067.pdb, with nucleotide G74. As shown in the attached png image (and PDB file for your verification), G74 is distorted too much from a planar purine base. 3DNA v2.1 (and DSSR) uses a geometric approach, not base name, to detect nucleotides, and G74 is beyond its very generous cutoff.

You may want to pay more attention to the case, or your MD results in general. Are those heavily distorted bases expected or they are a sign of some sorts of problem? If you insist, you can change the default cutoff by using the option: "-ntcutoff=0.3" or even higher.

HTH,

Xiang-Jun

[armanf]

Dear Xiang,
Thanks very much indeed.
I really appreciate it.
Just want to make a suggestion which I think might be helpful for MD analysis in general. It would be better if 3DNA had the option of skipping those pdb input files presented in the pdb_list and just gave a warning that for example this pdb contains distorted base-pairs and then resumed analysis of the rest of the files. Otherwise, one needs to  run it and checked any time for all the distorted files in a trajectory.

Thank you very much again and with the best regards
Arman

[xiangjun]

Hi Arman,

Just want to make a suggestion which I think might be helpful for MD analysis in general. It would be better if 3DNA had the option of skipping those pdb input files presented in the pdb_list and just gave a warning that for example this pdb contains distorted base-pairs and then resumed analysis of the rest of the files. Otherwise, one needs to  run it and checked any time for all the distorted files in a trajectory.

Thanks for the suggestion. I will see if I can find a simple solution. Presumably, the x3dna_ensemble script needs to check for the return value of an analyze run, and puts the problematic frame/structure into a file. I'll follow up with a new post along this thread.

Over the years, 3DNA/DSSR has benefited greatly from user feedback. I welcome any related questions on the Forum which always provide me a new angle of thinking.

Best regards,

Xiang-Jun
 
 

[xiangjun]

Hi Arman,

I've updated 3DNA to v2.2 in which the "x3dna_ensemble analyze" command contains the following new option:

Code: [Select]

--errlog, -g:   Print error message instead of abort program
Using your example of three frames (stored in folder pdbs/), I have the results shown below:

Code: [Select]

x3dna_ensemble analyze -b inp.inp -p 'pdbs/frame106*.pdb' -g

Code: [Select]

Process PDB file pdbs/frame1066.pdb 1 / 3
Process PDB file pdbs/frame1067.pdb 2 / 3
Error in running command: '/Users/xiangjun/X3DNA/bin/analyze -c  temp_model.inp 2> msgfile'
Process PDB file pdbs/frame1068.pdb 3 / 3
Please have a try and report back how it goes.

Xiang-Jun

[armanf]

Dear Xiang,
Thank you very much for your update. and sorry for my late late reply.
I have tried the new version and it seems that it works for me!
This will be a great help for those who carry out MD simulations.
Best
Arman