Wed Mar 12, 2014 3:50 pm by mleipold
Hi John,
There is some loss of sensitivity of the machine over very long runtimes (5-6+ hr). But we at the HIMC regularly run longer than that. In most cases, this slight drop is not sufficient to cause problems in your analysis. However, if you are running very large samples (several mL) due to particularly rare events, or barcoded samples, where you would combine several FCS files, this would be more of an issue than just running a lot of samples that are of comparatively short (<30min) acquisition time each.
1. Cleanliness of your samples is important. You will get less tuning drift if you do the water washes and then water resuspension to remove all possible free metal/antibody/buffer salts. Buildup of salts usually affects the Current parameter the most: you should have less than 1 Current unit drift over even very long runtimes (10+ hr) if you do the water washes, and this should basically keep you "on-peak". If you do not do the washes, or if you still have a dirty sample, then yes, you can shift yourself "off-peak" even in one day's runtime.
2. The 4 element calibration beads can allow you to renormalize your data to remove most of the decrease in sensitivity, before you combine FCS files.
3. In addition to Current-related salt build-up, according to DVS, the decrease in signal intensity supposedly comes from temperature-related changes in the TOF performance. Basically, your TOF start and Integration Time stay the same, but the actual ion peak shifts slightly to higher TOF. Therefore, you effectively lose some ion intensity because you're now missing part of your peak. As the machine gets hot, it will shift. However, when you shut down, it cools back down, so that it will have returned to the initial TOF time by startup the next day.
I have recently heard that some labs basically retune after 6 hr or so of runtime, including checking the appropriateness of the Mass 1 and Mass 2 TOF start settings. So far, this is not something that the HIMC has integrated into our user workflow....partly b/c I have grave misgivings about telling the average user to mess with the TOF windows.....
I think in most cases, with well-prepared samples that are <30min each, the 4-element beads are probably all that you would truly need for rigorous comparisons. But I have not formally tested this.
Mike