sci-chemistry
SHIFTX2 predicts both the backbone and side chain 1H, 13C and 15N chemical
shifts for proteins using their structural (PDB) coordinates as input.
SHIFTX2 combines ensemble machine learning methods with sequence
alignment-based methods to calculate protein chemical shifts for
backbone
and side chain atoms. SHIFTX2 has been trained on a carefully selected
set of
197 proteins and tested on a separate set of 61 proteins. Both the
training
and testing sets consisted of high resolution X-ray structures (less
2.1A)
with carefully verified chemical shifts assignments. SHIFTX2 is able to
attain
correlation coefficients between experimentally observed and predicted
backbone chemical shifts of 0.9800 (15N), 0.9959 (13CA), 0.9992 (13CB),
0.9676 (13CO), 0.9714 (1HN), 0.9744 (1HA) and RMS errors of 1.1169, 0.4412,
0.5163, 0.5330, 0.1711, and 0.1231 ppm, respectively. Comparisons to
other
chemical shift predictors using the same testing data set indicates that
SHIFTX2 is substantially more accurate (up to 26% better by
correlation
coefficient with an RMS error that is up to 3.3X smaller) than any other
program.
Please cite the following: Beomsoo Han, Yifeng Liu, Simon Ginzinger, and
David Wishart. (2011) SHIFTX2: significantly improved protein chemical
shift
prediction. Journal of Biomolecular NMR, Volume 50, Number 1, 43-57.
doi: 10.1007/s10858-011-9478-4.