Skip to main content
. Author manuscript; available in PMC: 2013 Apr 17.
Published in final edited form as: IEEE/ACM Trans Comput Biol Bioinform. 2011 Oct 17:10.1109/TCBB.2011.141. doi: 10.1109/TCBB.2011.141

TABLE 4.

Preservation of Structural Properties

Orig. Recon. Δ r m n

RMSD (Å) 1.937 (0.183) 1.938 (0.182) +0.05% 1.000 0.994 0.01
1.545 (0.183) 1.545 (0.182) +0.05% 1.000 0.997 0.01

Radius of gyration (Å) 11.92 (0.072) 11.92 (0.071) −0.01% 1.000 0.999 0.01
10.77 (0.075) 10.76 (0.075) −0.04% 1.000 0.998 0.02

SASA (Å2) 5083.9 (76.49) 5085.2 (77.03) +0.03% 0.997 1.004 −20.0
3863.8 (78.99) 3865.2 (78.87) +0.03% 0.998 0.997 13.8

φ RMSD (o) 19.36 (1.77) 19.33 (1.74) −0.12% 0.995 0.977 0.42
22.97 (4.52) 22.91 (4.50) −0.28% 0.999 0.996 0.03

ψ RMSD (o) 23.46 (3.18) 23.43 (3.18) −0.14% 0.997 0.997 0.04
23.97 (5.01) 23.89 (4.99) −0.35% 0.999 0.994 0.06

χ1 RMSD (o) 47.67 (7.12) 48.14 (7.0) +1.00% 0.987 0.969 1.95
54.57 (8.13) 54.98 (8.13) +0.75% 0.989 0.988 1.06

Average helical content 0.150 (0.015) 0.152 (0.015) +1.57% 0.860 0.844 0.03
0.231 (0.024) 0.235 (0.024) +1.75% 0.839 0.824 0.05

Fraction of native contacts 0.850 (0.028) 0.849 (0.028) −0.16% 0.912 0.908 0.08
0.885 (0.041) 0.886 (0.040) +0.13% 0.962 0.944 0.05

Average structural properties calculated before (“original”) and after compression/decompression (“reconstructed”) using PRIMO from trajectories of ubiquitin (first values) and protein G (second values). RMSD values are calculated for heavy atoms after superposition with respect to the first frame of the trajectory. Radius of gyration was calculated based on all atoms, including hydrogen atoms. Solvent-accessible surface area (SASA) calculations involved all heavy atoms and were performed with CHARMM[33]. Helical content was calculated based on backbone N(i)-H – O-C(i-4) distances of less than 2.6 Å. Native contacts were defined as residue separations of less than 4.2 Å for residues separated by at least 5 residues along the polypeptide chain in model 1 of the experimental NMR structure ensemble. Standard deviations are given in parentheses. Original and reconstructed values were compared with a linear regression analysis to obtain correlation coefficients (r), slope (m), and intercept (n).