One of the many recent innovations to make acquiring NMR data faster is the NOAH technique. NOAH concatenates several pulse sequences behind a single relaxation delay to reduce the total acquisition time. Various combinations of experiments are available, but perhaps the most useful is the NOAH_BSC variant which provides a gCOSY, 13C HSQC and 13C HMBC. In this post the NOAH_BSC spectra are compared with those collected using the standard Facility parameters.
To compare the NOAH_BSC spectra with standard spectra I used a 100 mg/ml sample of cholesteryl acetate in chloroform-d. The data were collected at 599 MHz using a 5 mm BBI room temperature probe. A standard NOAH_BSC parameter set from Bruker was used. The number of scans, t1 increments and sweep widths of the standard Facility experiments were adjusted to match the NOAH_BSC parameters so the spectral resolution should be the same. The table below lists some of the parameters used and the experimental times.
| Parameter | NOAH BSC | COSY | ASAP HSQC | ASAP HMBC |
|---|---|---|---|---|
| scans | 2 | 2 | 2 | 2 |
| increments | 256 | 256 | 256 | 256 |
| 1H width (ppm) | 15 | 15 | 15 | 15 |
| 13C width (ppm) | 220 | 220 | 220 | 220 |
| relaxation delay (ms) | 1500 | 1000 | 55 | 55 |
| acquisition time (ms) | 114 | 228 | 43 | 171 |
| NUS amount (%) | 100 | 100 | 50 | 100 |
| receiver gain | 9 | 18 | 2050 | 2050 |
| individual time (s) | 662 | 54 | 199 | |
| total time (s) | 1072 | 915 | ||
The total time for the individual experiments is about 15% less than the time taken for the NOAH experiment. This is due to the greatly reduced relaxation delay used in the ASAP HSQC and HMBC (55 ms), and the reduced acquisition time (43 ms) and use of 50% NUS in the ASAP HSQC. Newer NOAH experiments allow the use of NUS, and some include an ASAP element, so it may be possible to reduce the time taken for the NOAH experiment.
The signal to noise of the NOAH spectra seems to be significantly less than that of the standard spectra. The figure below shows the standard spectra (blue) overlaid with the NOAH spectra (red). For all three experiments the vertical scale of the NOAH spectrum was increased to approximate the intensity of the standard spectrum.
One possible explanation for the reduced sensitivity of the NOAH spectra is the use of a low value for the receiver gain in the NOAH experiment. This is probably required by the homonuclear COSY module, where the signal is large. In the individual experiments the receiver gain can be increased for the heteronuclear HSQC and HMBC to maximise the signal to noise. The individual experiments also have the advantage that the spectral widths can be set independently to increase resolution or reduce experimental acquisition time.
The NOAH_BSC experiment appears to offer spectra in a similar amount of time as the standard Facility experiments, however, it may be possible to optimise the NOAH parameters to reduce experimental time. More concerning though is the reduced sensitivity provided by the NOAH experiment. This comparison was done with a 233 mM sample. Most samples processed by the Facility are much less concentrated and sensitivity is much more of an issue. Close examination of the NOAH spectra also revealed that some correlations present in the standard spectra were missing from the NOAH spectra. At this point I would recommend sticking with the standard Facility parameters.
Acknowledgements
This comparison was suggested by Professor Bill Gerwick.
Hello,
ReplyDeleteI just chanced upon your blog while looking for something else -- there's so much great content & I honestly wish I had found it earlier! As someone who's been working on the NOAH experiments as part of my PhD at Oxford for the last few years, I felt inclined to comment here :P
I suspect that all of the SNR difference you're seeing is solely due to the RG. When we do SNR comparisons we always compare against standard library expts -- so for the HMBC our 'benchmark' would be hmbcetgpl2nd with all the parameters (d1, aq, rg, etc.) kept the same as in the NOAH run. [That's not to imply that the comparison you did here, against shortened individual expts, is 'unfair' in any way -- it's just a different comparison, in fact one that we've been intending to do for a while but never got round to...]
Anyway, for the NOAH HMBC, in my experience, we get ca. 20% sensitivity loss compared to the standard HMBC. And I'd expect that the ASAP HMBC, like all fast 2D techniques, will yield lower SNR than a standard HMBC (assuming otherwise identical parameters like rg). So, the SNR loss shown here in NOAH vs ASAP HMBC is almost certainly not intrinsic to the NOAH technique but rather, as you said, the consequence of setting RG so much lower.
Unfortunately customising that is out of our hands -- if we want different RG for different modules, the only way this can be done is to hardcode it in the pulse sequence, but clearly this isn't transferable to other samples/machines. We are definitely hoping that Bruker will fix it allowing us to do something like a per-module rg (and rga), but I'm not holding my breath either as I have to graduate soon :')
Actually, using more dilute samples would likely alleviate the RG mismatch, at least partially, because you could bump up the RG on the NOAH sequence. So the numbers would not look quite so bad (certainly not a 100-fold SNR difference). However, one can't go *too* dilute either because there's also a separate point (I saw this was raised in a previous post) about whether running the same NS for HMBC, HSQC, and COSY (in a NOAH expt) is an efficient use of time. That depends on the concentration -- if the sample is concentrated enough, then you'd always require NS=1 or 2 even if SNR is in abundance, so yes it is efficient. Most of our published comparisons implicitly assume this 'concentrated' regime (sometimes we call it 'resolution-limited' because the minimum expt duration is largely defined by TD1). However, this starts to fall apart for dilute samples, where (e.g.) COSY can be run with fewer scans than HMBC. Actually, some people argue that we should consider the NOAH as a "buy the least sensitive experiment (HMBC), and get the other experiments free (HSQC/COSY)", and I tend to agree with that. If the other expts are relatively cheap compared to the least sensitive one, as would be the case for a dilute sample, then you are not really gaining much from NOAH. Of course, what counts as "dilute" is also relative as it depends on the sensitivity of the probe/spectrometer.
Sorry for the essay -- I hope I'm not coming off as too defensive or anything! Just wanted to add a bit of (hopefully useful) insight for visitors etc - I'm more than happy to answer any questions too!
Cheers, Jon
(...as a means of identifying me, my name is on some of the more recent NOAH papers, and there is some contact info on https://nmr-genesis.co.uk)
PS - as of the time of writing, we don't run NOAH expts by default in Oxford either.