The filtration capacity of the mask is only one variable... the fitting and handling (Is the mouth-nose area completely sealed? Putting on/off, touching the face, etc.) is probably just as important, or more.
Don't get trapped into "X mask is better" and then don't follow up on the finer details.
It feels like one of the older papers, like penicillin, where you can read and understand the paper, and replicate the results if desired. You can also help the experiment by doing proper design. At least this one had a control.
My wife's professor keeps saying you don't need the positive or negative controls. (This is in a cytometry experiment, where you need to be able to tell if the flow is collecting what you expect, or not.)
And on behalf of statisticians and bioinformaticians everywhere:
"To consult the statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination. He can perhaps say what the experiment died of." Ronald Fisher, "Presidential Address to the First Indian Statistical Congress", 1938. Sankhya 4, 14-17.
Sort of depends what you mean. By definition you will be flow sorting cells on the basis of cell-surface markers specific to your fluorescent probes you apply to your sample. So, if you're flow sorting T-cells, for example, by definition anything that is picked up by anti-CD4 fluorescent Abs should be a T-cell. Of course, you should be validating the antibodies to make sure they actually are specific to the marker you're looking for, but that should happen in advance of any actual experiment.
Really, flow cytometry will give you a readout of the actual experiment you did to your cells of interest, but those would indeed need at least a negative control to establish if there is a difference between control and treatment groups.
You always need positive and negative controls -- that's what helps protect against oopsies or unexpected biology -- and how stuff like siRNA in c. elegans was discovered. In this case, it is a real experiment -- as in novel population, novel technique, and novel-to-flow-cytometry probes -- which really, really needs both controls. It isn't just collecting a well-understood population using standard probes and techniques.
If you're experimenting with new probes (since you are targeting a novel population) and new methods of marking (like using internal mRNA), you can get a lot of non-specific binding and/or inconsistent marking/fluorescence -- a lot of fluorescence works well on slides, but won't work for flow sorting. So, you can't really test without actually sorting, since the fluorescence for sorting is very different from that for a slide. If the sort depends on size and shape (to eliminate fragments and broken cells), the problem is even worse.
Even better would be to start with a known population and known probe, and confirm the technique works at all.
Last edited by Tod-13; 08-12-2020 at 09:42 AM.
Well, that would be validating your probes, no? That wouldn't really be a positive/negative experimental control. I was trained to always validate any antibody-based reagent in advance of any real experiments.
Negative controls sure - you should always have a sham/vehicle/etc group - but positive controls aren't always a realistic option. eg, if you wanted to test the hypothesis that Drug X will cause a leukocytosis if injected into mice, then you should definitely include a vehicle-only treatment group. A positive control however isn't really possible or desirable here.You always need positive and negative controls -- that's what helps protect against oopsies or unexpected biology -- and how stuff like siRNA in c. elegans was discovered.
You could inject a 3rd group of mice with bacteria to prove that they can indeed develop a leukocytosis, but what would the point of that? It shouldn't be surprising that wild-type mice can develop a measurable leukocytosis. Similarly, if you know that Drug X does accomplish other effects (let's say it leads to clonal expansion of cultured T-cells in vitro), you could use some Drug X from the same batch you gave your mice and give it to cells to show that the drug works. However, this doesn't really help you address hypothesis for the mouse experiment, either. There's also practical considerations like: how much drug X do I have/how many mice do I have/how much time and money do I have? And so forth.
A cleverly designed experiment in a well understood system should be able to take advantage of a positive control, but in practice I've found that it often doesn't work out that way.
There's nothing the matter with N=2. The error bars tend to be pretty big, though :-)
If you go out and measure the heights of people, and the first one is 5 ft 9 in, and the second is 6 ft 1 in, you now know more about the likely distribution of heights than before you measured the first one.
People have funny ideas about statistics, e.g. that P=0.04 and P=0.06 are hugely different in a true/false sense. In reality, they are really close.