Table 1 Parameters used for benchmarkinga

From: Solution-state methyl NMR spectroscopy of large non-deuterated proteins enabled by deep neural networks

Parameter

HDAC-like spectra

MSG-like spectra

Number of signals

275

600

Larmor Frequency (MHz)

\(\in (600,700,800,950)\)

\(\in (600,700,800,950)\)

1H SW (Hz)

2000–5000

2000–5000

13C SW (Hz)

2000–000

2000–5000

1JCC (Hz)

34 (2)

34 (2)

\({R}_{2}^{\left(1\right)}\) (s−1) (1H dim)

50 (10)

60 (15)

\({R}_{2}^{\left(2\right)}\)(s−1) (13C dim)

50 (10)

60 (10)

  1. a Where values are given as a range, for each spectrum the true value is taken as a random value from a uniform distribution of the range. Where values are given as a number followed by a bracketed number, for each signal the value used is randomly chosen from a normal distribution centred on the first number and with a standard deviation given by the bracketed value. For all synthetic spectra, half of the peaks are chosen as resulting from a methyl moiety, while the other half come from non-methyl 13C-1H moieties, i.e., leading to triplets in the 13C dimension.
Back to article page