LC/ single MS Quantification

[klimbaka]

Hello everybody, why can't I find a topic that discusses LC/MS quantification using a single stage MS. I am working with hormones and so far I have obtained curves with a stable slope and a good R squared, and thus good reproducibility. My MS is an ion trap.

[Kostas Petritis]

As your results indicate is possible, especially with the latest generation ion-traps that have much better AGC (or AGC type algorithms/devices). Single quadrupoles should still be at least somewhat better than ion-traps when it comes to quantitation...

[TobiWms]

Quantitation on Ion Traps will vary widely depending on the manufacturer. Here's how Varian does it on the 500-MS with linear range of 10'4, r2= .09998, and negligible mass shift.
http://www.traceorganic.com/2006/presen ... lt.ppt.pdf

[sindhu]

When you work with hormones you really need a very high sensitive MS system. You can use either ABI 4000 or ABI 5000 depending on your budget availablity.

With regard to usage of MS/MS only for your application is because of MRM which you get only from Triple Quad systems and from Sindle Quad systems.

Basically single quad systems are best used for qualitatative applications and it cant be used for quantitation because of non availablity of collision cell & 2nd quadrapole.

By using CID you can get some qualitatative information on fragmentation but its not quantitative.

By advice is only triple quads with high sensititivity are best recommended for analysing harmones.

[sassman]

For quantitative work, you will get best results from a triple quad in MRM mode. That doesn't mean that you cannot do quantitative work with an ion trap. My experience has been that ion trap instruments do not have as wide of a linear range, cannot handle very dirty samples, and are not as stable as a triple quad. A trap is, however, often better for qualitative work.

Most people use MS/MS for quantitative work because it is much more suited to real (dirty) samples than single stage MS. When using single stage MS, any compound with the same mass as your analyte will interfere with your analysis. If you use MS/MS, something must have the same precursor AND product ion mass in order to cause an interference. The result is a much more sensitive analysis.

You might be getting nice calibration curves with standards, but I suspect you will have many problems when you try to run real samples in single MS mode. BTW, you should be using internal standards (preferably isotopic) to correct for matrix effects.

[dr_Pyrex]

I can add to this discussion using labelled internal standards. It's make quantitation much better, especially in residue analysis in complex matrices like tissues or environmental samples.
We are working with residues of veterinary drugs in food with LC-MS, MS/MS (ion trap and triple-quad). We use multi-residue methods and we always tried to have as many labelled IS as possible. Without it quantiation is often impossible.
For hormones analysis you can find labelled IS for every analyte.

[lmh]

I'm going to have a bit of a rant here, so please forgive me:

Properly used, single quad MS systems are perfectly OK for quantification. Of course they have less selectivity than a triple quad, but they have vastly more selectivity than an HPLC-UV system, and there are plenty of HPLC methods in use around the world!

Nevertheless, reviewers who wouldn't bat an eyelid at an hplc method routinely turn their noses up at single quad methods on the grounds they are not as selective as triple quads. This is highly aggravating behaviour!

Of course a triple quad system can do MRM, and MRM can be better than single ion monitoring. But if you use a trap that is designed to activate only the parent ion and not the fragments, it is absolutely vital that you choose your transition with care! There are, unfortunately, plenty of people setting up MRM methods using silly transitions such as loss of 18, which is possible for any molecule with a hydroxyl group. This renders the method barely better than single ion monitoring.

Sensitivity: this is a sore point. Sensitivity is usually best measured as S/N ratio, but this ratio is only accurate if both S and N can be quantified accurately. Since N is very nearly zero in MRM methods, the S/N ratio looks enormous. But the measurement is dreadful. To take an extreme case, if you work with so little sample that the signal is a single count, but there is no noise, the S/N ratio is still infinite, even though you are absolutely at the detection limit of the instrument, and your next replicate injection might very well have an error of 100%. More realistically, it's very easy to get to a situation in MRM where the shape of the peak in chromatography is ragged and horrible, because the signal is really very small, but the S/N ratio is vast. Of course the creative analyst simply sticks a well-chosen Gaussian smooth function through the spike and gets a beautiful Gaussian peak...

Trap sensitivity is limited by the trap's capacity. Frequently SIM and SRM do not increase the actual measured signal (depending on design of instrument). Quad-based systems, whether single or triple, actually generate a bigger signal if used in SIM or SRM. Thus a simple, cheap single quad can sometimes out-perform a trap on sensitivity (in terms of limit of quantification based on error-term in the calibration curve points), even though it has a worse S/N ratio.

[/i][/u]