[arnab]

Hi
I recently did a enzyme-inhibition experiment for my unit 6 coursework and i am now on the part where i have to carry out a statistical test on the data i collected from my experiment. This is what my experiment was bascally about :


It was just a simple enzyme-inhibition experiment. SO in a simplifed version :

1. I first poured 5cm3 of milk in to 4different test tubes
2. To another set of 4test tube, I then added 2cm3 of trypsin to each of them and then adden 1cm3 of the inhibitor at a concentration 10-4 mol dm^3.
3. I then poured the solution containing the enzyme trypsin and inhibitors into the milk
4. I then took reading for the Absorbance value from the colorimeter, at a regular interval for 15second for the next 4minutes.
5. I then repeated the experiment but with different concentration of the inhibitor, i.e 10^-3 mol dm^3 & 10^-2 mol dm^3

Now i was thinking of using the spearman'r rank correlation Coefficient to determine a relationship between the absorbance value for different inhibitor concentration but i am unsure as to how to go about doing it.

Am i right in understanding that i should i just use Absorbance values I collected for the lowest and highest concentration of inhibitor in order to determine there is a relationship between the different concentration of inhibition?

Or am i completely wrong in using this statistical test? PLEASE ADVICE!

[arnab]

anyone?

[Waterstorm]

I kinda skimmed what you wrote, but a spearman's rank correlation cofficient will only show how strong or weak your correlation is - ranging from a valve of -1 (perfectly negative) though 0 (no correlation at all), to +1 (perfectly positive). It won't show anything else.

If you're trying to compare the your two variables (the x and y values on a scatter graph) you can use a linear regression model to work out of there if there is any relationship between x and y - i.e. does x actually affect y, or are does it not make a difference at all. You can also, from this, work out a confidence interval to show how reliable your hypothesis is.

If that makes sense though. What is it exactly you want to show with your statistical analysis? From that I could give a good test to use.

[Waterstorm]

Actually for the linear regression model you gotta assume a normal distribution for your data, so dunno if that'll work out.


For a correlation coefficient, do you have all your x and y values? Cause that's all you need.

[arnab]

(Original post by Waterstorm)
Actually for the linear regression model you gotta assume a normal distribution for your data, so dunno if that'll work out.


For a correlation coefficient, do you have all your x and y values? Cause that's all you need.

Yep, i got my values of x and y. I like prove the relationship that as x and y increases, the rates of reaction also increases.

E.g say X = 1% concentration of a solution
Y = 2% concentration of another solution

So like to prove that as X and Y increases, the rates of reaction also increases.

[Waterstorm]

Do you know how to work out the Pearson's correlation coefficient? The only difference is, is that Spearman's isn't affected by extreme values.

And as long as you have the x and y values, it's just a case of ranking x from lowest to highest with their corresponding y values, then working out the difference between the two. And then using the equation. Tbh, I never really liked this, and didn't understand it fully. I can message you our uni notes on it though, it's explained really well on there, if you want.