On Fri, 11 Aug 2006 05:59:49 +1000, "Dexter Chun" <dexter81 at gmail.com>
wrote:
>Hi all,
>>I'm trying to figure out the basis behind the 'right' calculation for viral
>titre (u/ml) based on serial dilution of viral supernatant on NIH 3T3 cells.
>My transducing efficiency is assessed based on eGFP expression by
>FACS. Hence the graph is constructed using %eGFP v serial dilution fold.
>I've been told to choose the point on the best linear part of the curve, *
>intuitively* the midpoint of the straight portion to get my titre (u/ml).
>Is this the right or best method?
>>What is the right way to go about the graphing? By using all the data
>points, I have a 'weird' curve. Is the viral titration meant to result in a
>linear correlation theoretically?
>a) do i remove the extreme point, and plot a linear curve based on remaining
>ones
>b) or is there some mathematical conversion of points OR the axes
>
In principle, it should be linear -- at least until it saturates.
If it is not, you need to do some troubleshooting. What does "weird"
mean? (That is not an officially recognized term for describing
curves. :-) ) Does it have some shape, which might be a clue? Or is it
just erratic?? What is reproducibility, for various steps?
When you say %GFP, you mean % of cells showing it?? (So that level of
expression per cell is not an issue.)
What do you need from the assay? Sometimes, an assay is "useful"
within some range, even though it is not behaving properly. That is,
it may be a reproducible lab measurement, of uncertain meaning. But if
twice the score means twice as many cells, within some limits,
sometimes that is useful, even if you are not sure whether it went
from 1% to 2% or 0.1% to 0.2%.
bob
