We are having reproducibility problems with our Quantiva MS. Absolute signal response is varying by 30-40% for replicate injections of dansylated neat estrogen metabolite standards at 25 pg/ml. This improved to about 5-6% when corrected for labeled C13 internal standards but should be closer to 1%. We also see "cross-talk" of up to 1% at 3 amu away. Please share your experiences.
Hi Neal, I am not sure what is the maximum accepted percentage for the variation in reproducibility, but let me tell you what I know.
It seems that you are facing a matrix effect, because when you used an isotope internal standard, you were able to decrease the variation significantly. If this is a case of a strong matrix effect, then you may need to study it and come up with solid findings before you accept or reject a result from the MS.
Another problem you are facing is the shift in the mass of the parent or products ions by 1%. Put it this way, in the MS software you have the ability to setup your mass scan window, which we used to set it as 0.2 amu as recommended by the service engineer. This means, if the mass of the parent ion was 123.4, it scans from 123.3 to 123.5. In other words, we are allowing for shift in the mass of any ion by +- 0.1 amu. Clearly we are not making the allowance in percentage as you do. But, I see that 3 amu is a very big allowance, you may need to investigate this value again.
One more thing, a good instrument should give a good reproducibility for duplicate injections. You can check this by calculating the CV (correlation of variance) between the duplicates. Refer to the paper attached for some examples.
Mukhtar,
Thank-you for your comments. I pose this question because we believe that the instrument is not performing at specifications. To your first response, the data is from neat standards so we have not added the issue of matrix effects. In real samples, the issue is much more complicated and matrix effects do factor in.
Previously, we have reduced the window to 0.2 amu in order to be more selective and reduce "cross-talk". There is some improvement but there is a greater decrease in sensitivity. With a window of 0.7, we should not see 1% in windows 3 amu away. The instrument has just been serviced, cleaned and calibrated by Thermo. It should be at top performance.
The signal reproducibility should always be better when corrected for a labeled internal standard. We would expect absolute signal %RSD of neat standards to be
One problem solved. The Divert Valve was faulty and did not return to the same position causing variable amounts of sample to reach the MS. Still trying to understand the cross-talk. It does not occur for vitamin D, retinol, or caffeine, but it present when dansyl chloride is used.
Gaeten,
Thank-you. The estrogen metabolites are labeled and the dansyl derivative is formed to get them to ionize. The dansyl fragment is monitored and there is no C13 label at that point. The difference in mass is used a Q1 to select for the parent ions + dansyl. Your explanation makes sense. We knew that the natural abundance of C13 would not explain the cross-talk but we are only monitoring the dansyl at that time. A 4% natural abundance of S34 explains why we see it across all metabolites.
- Badges
- Science topic