Skip to main content
. Author manuscript; available in PMC: 2016 Sep 1.
Published in final edited form as: J Biomol NMR. 2015 Jul 4;63(1):39–52. doi: 10.1007/s10858-015-9961-4

Table 2.

Prediction results for different atom typesa

Hydrogen n trim xrms sdev Carbon n trim xrms sdev
AH2 911 24 0.20 0.41 AC2 569 6 0.61 1.02
AH8 903 23 0.16 0.25 AC8 557 7 0.68 1.09
GH8 1317 45 0.14 0.34 GC8 767 18 0.69 1.61
CH5 1148 22 0.11 0.26 CC5 597 21 0.54 0.68
UH5 854 19 0.14 0.28 UC5 461 13 0.74 1.04
CH6 1154 22 0.10 0.19 CC6 649 15 0.59 0.95
UH6 863 18 0.11 0.17 UC6 490 10 0.81 1.19
AH1′ 905 20 0.14 0.19 AC1′ 527 10 1.03 1.29
GH1′ 1321 26 0.12 0.30 GC1′ 694 10 0.93 1.31
CH1′ 1143 14 0.12 0.18 CC1′ 605 17 0.63 1.07
UH1′ 845 20 0.14 0.19 UC1′ 471 13 0.92 1.42
AH2′ 834 15 0.15 0.19 AC2′ 352 11 0.72 0.80
GH2′ 1215 32 0.12 0.21 GC2′ 508 18 0.67 0.76
CH2′ 1034 29 0.10 0.19 CC2′ 467 12 0.48 0.87
UH2′ 779 15 0.13 0.21 UC2′ 361 10 0.45 0.56
AH3′ 767 18 0.14 0.19 AC3′ 334 8 1.37 1.71
GH3′ 1125 31 0.13 0.23 GC3′ 468 14 1.16 1.65
CH3′ 959 23 0.11 0.15 CC3′ 434 11 1.07 2.02
UH3′ 697 6 0.11 0.15 UC3′ 331 1 1.56 2.16
Total 18,774 422 0.13 1.33 Total 9642 225 0.83 28.44
a

Output from the support vector regression analysis. The columns labeled n represent the number of shifts of each atom type used in the SVR regression analysis. The columns labeled trim represent the number of shifts that were trimmed out between the first and second SVR steps. The columns labeled xrms represent the square root of the mean of squared deviation between predicted and measured chemical shifts as calculated in the tenfold cross-validation. The column labeled sdev is the standard deviation of all the experimental hydrogen or carbon shifts for the corresponding atom type