Dear Ouyang,
I am not familiar with Material Studio4.3 you used to generate the RNA structures. However, thanks to your attached PDB file, I could check why 3DNA did not work. As is clear from the RasMol rendered image (attached below), your RNA structure is heavily distorted. To make my point even clearer, I extracted base-pair A51 vs U33 (coordinates attached) and rendered it with RasMol (also attached below): note the extra-erroneous connections in A, and the non-planar geometry of U. Moreover, the distance of N6(A)...O4(U) is 4.9 A, and N1(A)...N3(U) is 5.6 A.
Based on my experience, I would hardly take this as a valid pair at all. Still, if you define A51 and U33 in 'baselist.dat' and set H-bond distance cut-off to 5.0 A instead of the default 4.0 A in 'misc_3dna.par'
# Section 1: parameters defining H-bonding, in angstrom
# primary H-bond
<hb_dist1>5.0</hb_dist1>
A51 and U33 will still be identified as a pair:
1 2 1 # 1 + ....>-:...0_:[A51]a-**--u[U33]:...0_:-<.... 3.47 1.12 43.35 10.92 7.88
##### Base-pair criteria used: 5.00 0.00 15.00 2.50 65.00 4.50 7.50 [ O N]
HTH,
Xiang-Jun
[attachment=2:1psgyjp7]ARNA-ms.png[/attachment:1psgyjp7]
[attachment=1:1psgyjp7]ARNA-bp1.png[/attachment:1psgyjp7][attachment=0:1psgyjp7]ARNA-bp1.pdb[/attachment:1psgyjp7]