GUIDELINES FOR THE USE OF OVERSAMPLING IN PROTEIN NMR

With detailed NMR studies being pursued on ever larger proteins, the question arises as to whether the typical 16-bit analog-to-digital converter (ADC) has sufficient dynamic range to digitize the sum of signals as well as a small signal of interest, even in the presence of excellent solvent suppression. Oversampling may sometimes be indicated to alleviate this problem. The noise in an NMR spectrum consists of analog receiver noise, analog noise from the analog-to-digital converter components, and ''quantization noise'' due to the limited precision of vertical digitization. Oversampling (i.e., sampling faster than required by the sampling theorem) may decrease the influence of the latter two noise components, both of which have high-frequency components and are not influenced by audio filters. In addition, oversampling in the indirect time domain may be of value under certain conditions. Signal-to-noise ratios were evaluated for doped and undoped H2O/D2O mixtures over a range of receiver gains for 1D and 2D proton NOESYs with or without oversampling in direct and indirect time domains. The results suggest general recommendations concerning the use of oversampling. These general principles allow spectroscopists to rapidly determine whether oversampling in direct or indirect dimensions will be useful under their particular experimental conditions. (C) 1995 Academic Press. Inc.