NUPARM vs X3DNA twist values

[kailsen]

Hi,
 I did some step parameters calculation using 3DNA as well as the NUPARM web server present in the following URL
http://http://nucleix.mbu.iisc.ernet.in/nuparm/nuparm.shtml
The structure is a RNA with GU mismatch in the 4th n 5th steps. The pdb id is 2G3S(attached the coordinate file).
The dinucleotide step contributed by non Watson-Crick base steps alone seems to have different values for Twist (x3dnaTwist=49.67, NUPARM=24.20). On visually examining the step the does not show twist value as high as ~50 degrees. I am aware that the geometric calculation involved in both these programs are different but all other values correlate well including the base step, intrabase step and torsion angles values. Do u think shearing in GU base pair seems to be a reason But, it would be good if u can explain the difference occuring in twist value alone. I have not compared the values with CURVES+ program yet   .Using SCHNAaP/CEHS the twist value is 25.52 which is  close to NUPARM values.

[xiangjun]

As illustrated in Figure 3 of the standard reference frame paper [J. Mol. Biol. 313(1), 229-237, 2001], and in the section "Treatment of non-Watson±Crick base pairing motifs" of the 3DNA 2003 NAR paper (also Figure 3), the discrepancy you noticed in Twist angle between 3DNA and NUPARM is due to large Shear of the G-U Wobble pairs.

You may need to (re)read the two papers for further details. It would be helpful if you repeat your calculations with Curves+ and post back your results: I would guess that Curves+ twist angle be pretty similar to that from 3DNA.

HTH,

Xiang-Jun

PS: Please also see my blog post "How shear affects twist angle of a dinucleotide step?"

[kailsen]

Hi Lu
Thanks for the info.. As  you said, i just had to re-read these earlier papers carefully once again. and understanding the reason was not much of a diff task. Thanx for pointng me in the right direction. BTW, i could nt still CURVES+ program. having problems in running it. Using it for the first time !!!. Asi get the values will post it.