Hi Everyone!
How much of a between-sample variance in retention time would you accept in a targeted proteomics assay (when analysing similar samples)?
I'm assuming you are referring to reversed phase HPLC reproducibility. If the HPLC system has been allowed to fully equilibrated at the initial solvent conditions you should be able to achieve better than 0.1% reproducibility. This implies the baseline has returned to the level of the previous/initial injection. However, if there is non-specific binding occurring due to impurities bound to your stationary phase then all bets are off. A periodic solvent cleaning of the separation column according to prescribed manufacturer's methods maybe necessary to ensure peak performance. Also your pumping system has to be reproducible; so a check of flow rate reproducibility may be warranted if you are not achieving the above 0.1%.
the retention time reproducibility can be effected by many factors such as what mentioned in the previous answer , and especially if you employ gradient method .
any way if you unable to overcome this situation by good washing , maintain column temperature , pump and LPG calibration .
if all the solution are failed to solve the reproducibility , then my first choice is spiking method .
if you do not know the spiking method message me and i will discuss it with you .
best regards .......
Around +/-2 min. Is your nano LC system split flow or split-less? A split-less one would deliver more reproducible gradients between runs. Do you observe a systematic RT shift for all transitions or just a particular one?
Hi Liyan! With one column type I get less than 30 sec. of shift with an unfractioned cell lysate as sample, using a 30 minute gradient. I am relatively comfortable with this. But another column gives a larger shift (for many of my targets), but has better sensitivity. As an "LC person" I feel like disqualifying the latter column (for a column comparison study), as I consider RT to be factor for identifying a compound. But, I am curious to what the rest of the community thinks.
It highly depends on the hydrophobicity of the peptides. A recent paper
(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3918884/) on iRT showed an average CV of 2.5% over 9 runs.
Hi Anthonius! The shifts are not systematic. However, 10-30 second shifts is (just) under 2 % variation in RT for the target peaks; This is just within acceptable according to several guidelines for LC-MS analysis of therapeutics etc. Therefore, when another column type spreads wider than this, I would feel that this is no good for identification-wise. But what does the proteomics community think? e.g. Liyan feels that a 4 minute window is common. But how much of an variation do you think is acceptable for a 30 minute gradient of a lysate sample?
On identification: If you monitor 2-3 peptides per protein X 2-3 fragments for each peptide, and they all show the same up/down-regulation trend across your samples, then you can be confident of the results. You can further validate the results by setting the MS to perform product ion scans after MRM trigger or re-inject the samples on the same system for product ion scans.
Now that you have mentioned that your run time is only 30 min, +/- 2 min shift is on the large side. Maybe you can try stretching your gradient to 45 or 60 min and/or run peptide RT standards on the column to troubleshoot.
- Badges
- Science method