Obviously, the build-up of the two direct NOEs is only proportional to their g values for tm=0, at the very beginning of the curves, which then deviate more and more from a straight line to reach some sort of plateau for very long tm values. This initial slope of the NOE build-up curve corrsponds to the cross-relaxation rate o, which depends on the interproton distance It can be measured in two ways by measuring 2D NOESYS with a series of increasing tm delays(e. g, at 50, 100, 200, 300 ms), and then fitting an exponential function to the data points extrapolating down to t =0. This will give very reliable results, but require a lot of spectrometer(and processing and evaluation time) or by measuring just a single 2D NOESY experiment with a sufficiently short tm, so that the build-up curves can still be considered linear in good approximation(in our example, with tm< 100 ms ). This will be much easier than actually tracking the whole build-up curve. The problem, however, lies in the term "sufficiently short". If one chooses tm too short, then the NOE cross-peak intensities will be just too low to measure; if tm is not short enough, then the measured intensities will show considerable deviations from the g ratios! Solution either measure the whole build-up curve once and assume that it will be similar for all molecules with similar tc (i.e, similar molecular weight at similar temperature and solvent ). Or: take an educated guess (e.g, from tm values given in the literature) ndirect NOEs can be recognized by their initial slope which is zero! This means that they only show up at longer mixing times. However, when a"sufficiently short "mixing time is chosen still too long they can be mistaken for"real "direct NOEs If all these precautions are kept in mind, NOEs (cross-peak integrals) can be converted into interproton distances. This is usually be done with the help of a reference NOE corresponding to a known distance(e.g, CH2 groups, H-C-C-H fragments in aromatic rings etc.)102 Obviously, the build-up of the two direct NOEs is only proportional to their s values for tm =0 , at the very beginning of the curves, which then deviate more and more from a straight line to reach some sort of plateau for very long tm values. This initial slope of the NOE build-up curve corrsponds to the cross-relaxation rate s, which depends on the interproton distance! It can be measured in two ways: - by measuring 2D NOESYs with a series of increasing tm delays (e.g., at 50, 100, 200, 300 ms), and then fitting an exponential function to the data points extrapolating down to tm =0. This will give very reliable results, but require a lot of spectrometer (and processing and evaluation time). - or by measuring just a single 2D NOESY experiment with a sufficiently short tm , so that the build-up curves can still be considered linear in good approximation (in our example, with tm £ 100 ms). This will be much easier than actually tracking the whole build-up curve. The problem, however, lies in the term "sufficiently short". If one chooses tm too short, then the NOE cross-peak intensities will be just too low to measure; if tm is not short enough, then the measured intensities will show considerable deviations from the s ratios! Solution: either measure the whole build-up curve once and assume that it will be similar for all molecules with similar tc (i.e., similar molecular weight at similar temperature and solvent). Or: take an educated guess (e.g., from tm values given in the literature)… Indirect NOEs can be recognized by their initial slope, which is zero! This means that they only show up at longer mixing times. However, when a "sufficiently short" mixing time is chosen still too long, they can be mistaken for "real" direct NOEs. If all these precautions are kept in mind, NOEs (=cross-peak integrals) can be converted into interproton distances. This is usually be done with the help of a reference NOE corresponding to a known distance (e.g., CH2 groups, H-C-C-H fragments in aromatic rings etc.): r r I I ij ref ref ij = 6