question on reference frame (ref_frames.dat)

[kbt22]

Hello,

First let me say that your program 3DNA is a wonderful gift to the scientific community!

I have a question about the reference frame used in your program, 3DNA.  

First, let me tell you a bit about my problem.  I am specifically interested in examining the Lac-R repressor and its structure when bound to known operator sequences.  The two segments of DNA contained in the .pdb file (1LBG.pdb) are largely distorted by the binding and appear to have a near 45 degree kink.  

My question is regarding the ref_frames.dat file.  When I look at my .pdb file and its interpretation of the DNA helix visually and overlay the ref frame returned by your program, the reference frame in the ref_frames.dat file seems to be contradictory to the helical structure.  Let me explain further.  While the coordinates returned in the ref_frames.dat file do appear to be in the "center" of the base pair, the reference frame gives a z-axis that does not appear to align with that of the DNA.  Also, the x-axis, which you say is supposed to point away from the minor groove, appears to be pointing in some other direction!  

My question is this: is the frame given in the ref_frames.dat file, the same as the one you refer to as the coordinate frame, where the positive x-axis is defined as pointing away from the minor groove?  If so, why do the axes appear to be pointing in these strange directions?  

I hope these inquiries make sense.  I was going to attach the .pdb, .out, and .dat files, but I cannot figure out how to do it!  If you care to reproduce my scenario, the pdb file is called "1LBG.pdb".  I simply ran

Code: [Select]

find_pair 1LBG.pdb stdout | analyze

Thank you in advance for your help, and I hope I have not carelessly missed something in the User's Manual!

Kevin Towles

[xiangjun]

Hi Kevin,

Thanks for using 3DNA and for your kind words regarding the 3DNA software! It has always been a pleasure to do something meaningful to myself and helpful to the community. Over the years, it is users like you that help move the project forward. So I always welcome any (technical) questions regarding various aspects of 3DNA.

Now back to your questions. Two issues need to be clarified.

Firstly, the base-pair reference frames as output from 3DNA are indeed as advertised for Watson-Crick base-pairs where the minor groove side of the two bases in a pair aligns (facing the same direction). Specifically, in such cases, x-axis points from minor groove side to the major groove side, and y-axis points to the leading strand, and z-axis completes a right-handed coordinate system and would point in the 5' to 3' direction of the leading strand. As always, this is best illustrated with an example, and I am using bdl084.pdb, which is distributed with 3DNA. Try to reproduce the following:

Code: [Select]

find_pair bdl084.pdb bdl084.inp
frame_mol -f ref_frames.dat frame.pdb frame.alc

block_atom bdl084.pdb
comb_str bdl084.alc frame.alc bdl084_frame.alc

rasmol -alchemy -noconnect bdl084_frame.alc

Please note that there is a bug w.r.t. displaying Alchemy file in RasMol 2.7.x. I've been using v2.6.4, the version directly from Roger Sayle, the original author of RasMol.

Secondly, when a pair is no-longer of Watson-Crick type, the minor groove side of the two bases in a pair could be in quite different relative orientation. In such cases, where would the minor groove side of the pair be? And this is where you noticed "the axes appear to be pointing in these strange directions". Check carefully the base-pair parameters from the 3DNA main output file, and you would see the point. The "strange directions" are actually not that strange in that they are also the "middle" base-pair frame, as in the Watson-Crick pairs.

HTH,

Xiang-Jun

[kbt22]

Thanks for the speedy reply!

Well, I checked the example problem that you cited, and it appears to be reasonably similar to what I am seeing.  According to your reply, the ref_frames.dat file is exactly what I thought it was.  Thank you for your help,

Kevin