A/B/Z forms

[auffinger]

Dear Xiang-Jun,

Thanks a lot for all the modifications you have done on request to your programs.
It is great that dssr provides now clues about the helix form for each base pair (and nucleotide).

Unfortunately, I think that the manner its added to the output files, although quite informative, is also
very difficult to parse. Could you imagine adding these A/B/Z labels to, for example, to
the torsion files (for dssr and also for 3DNA - this would be very useful to us
and hopefully to others too).

For example here for the dssr torsion file

        nt             bin    cluster   suiteness  A/B/Z/
 1     ..A.C.1.        inc      __       0.000      xxx
 2     ..A.DC.2.       33p      1a       0.824    xxx
 3     ..A.DG.3.       33p      1a       0.403    xxx
 4     ..A.DG.4.       33p      1a       0.387    xxx

and here for the 3DNA torsion file

              base      chi A/S     alpha    beta   gamma   delta  epsilon   zeta     e-z BI/BII     A/B/Z
   1 A:...1_:[..C]C  -161.7 anti     ---     ---     59.1    78.4  -155.0   -71.6   -83.4  BI         xxx
   2 A:...2_:[.DC]C  -163.8 anti    -60.6   156.1    55.3    82.7  -175.3   -64.1  -111.2  BI    xxx

Cheers,

Pascal

[xiangjun]

Hi Pascal,

Thanks for your feedback to the new functionality of DSSR. To clarify, please note the following two points:

• The Richardson et al. sugar-phosphate backbone suite name classification deals with a single-stranded dinucleotide (e.g., ApG),  and the 53 conformers have been derived primarily from RNA, but not DNA. Try the classic B-DNA Dickerson dodecamer (CGCGAATTCGCG, 355d) as an example, you'll see many outliers ('!!'). I implemented the suite classification in DSSR mostly for a better understanding of this important community consensus on RNA backbone. You may well want to read the original Richardson et al. (2008) paper and use the official Suitename program.
• The DSSR A-, B-, Z-form classification method works on a double-stranded dinucleotide step (e.g., ApG/CpT, not a base pair). It is different from the algorithm implemented in 3DNA, and is still experimental.

Thus, I cannot think of a consistent and meaningful way to implement your two suggestions in DSSR. If you find the current DSSR A-, B- and Z-form classification useful in your project, you are welcome to employ the info as you see fit.

Xiang-Jun

[auffinger]

Hi Xiang-Jun,

I understand your concerns. For me it is more related to having this
info somewhere in your dssr output files for easy parsing where ever the
right place might be.

As it is it's just very difficult to parse.

You say that the algorithm used is different from that in 3DNA.
Could you comment on that ?

Thanks,

Pascal