1
$\begingroup$

I've collected fluorescence data from some bacterial cells.

Each cell has a gene in it which can be induced to fluoresce. However, even without being induced, the gene will still fluoresce a little.

I have subjected the cells to a series of concentrations of inducer, including 0% inducer, and measured the resultant fluorescence over time.

The fluorescence produced from cells that have not been induced is, effectively, 'background' fluorescence. Therefore, I would like to subtract the background fluorescence from the fluorescence measured in cells that have been induced in order to determine the fluorescence that is caused by induction.

$n = 3$ for all measurements.

My question is, can I take the mean fluorescence of the uninduced cells and assume that it is the true mean? I.e. assume that its $SE = 0$. If I can do that, when I subsequently subtract the mean background fluorescence from the mean induced fluorescence, the final calculated $SE$ will be lower.

Here's some example data:

 Inducer:       0%              5%
 1             3411            4965
 2             3427            4784
 3             3412            5379
 Mean          3416.666667     5042.666667
 SD            7.318166133     249.0385958

Given that the background will change for each experiment, I assume that the mean background measured is the true mean for that experiment. Is that not correct?

I must state that my maths/stats skills are rudimentary at best so please bear that in mind when answering.

$\endgroup$
0
$\begingroup$

No, you can't because the variance of the error term will be underestimated as a result. The simplest way to analyze your dataset is a two sample t-test where variances are not considered equal because they are clearly unequal for your data.The formula can be found here:

http://www.itl.nist.gov/div898/handbook/eda/section3/eda353.htm

On the other hand, your proposition is similar to one-sample t-test, which is not the case here because the 0% mean is estimated rather than known.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.