When normalizing to control I usually divide the treatment reading with the reading for the control. But how do we deal with this when the control value is 0? For now, I'm just plotting absolute values but since my data is from in vitro studies it would really help if I could normalize to individual assay control.
hi
I am not at all in this field but the q seems of a general mathematical nature so a possible suggestion from electrical engineering is to use per unit method..that is divide each data by its base unit so control vector by its base(eg max value .) and the treatment vector by its max value for comparison..then both vectors can be compared on pu basis. Hope it helps
Cheers
If this is a control signal input, you can use the standard deviation value to divide each data by its base.
Thanks Narasim & Utkal! Narasim, just to clarify since I'm not aware of what you refer to as 'pu basis': I would have to divide my control values by the control max or min value and the treatment values by treatment max or min values is that right? Would this method still work if the variability in my data was high? One of the main reasons I wanted to normalize to control was because the variability between assays tends to be high at times..
Utkal: Do you mean the standard deviation value of the control? I'm sorry, I'm a bit confused, I've never done this before - would u please mind clarifying?
hi
It might help to obtain useful answers if a small representative data is presented .However see the following.
q:I would have to divide my control values by the control max or min value and the treatment values by treatment max or min values is that right?
ans .That is correct.
q:Would this method still work if the variability in my data was high?
Ans : It is important I think to choose the base depending on the problem .Please see attached pdf.
hope it helps
Cheers
You could also consider making the zero values equivalent to the lower detection range of your assay.
Hello to all, i also want to normalize some nonlinear variables with respect to linear. The Linear are constant,e.g. If i want to normalize A_Nonlinear with respect to A_linear, which is constant i can try: A_nonlinear/A_linear. The problem i am facing is that the some of the linear are zero, but nonlinear start from zero and varies. So, i don't know how to normalize it. My objective is to demonstrate the variation in nonlinear with respect to linear. Any help??
- Badges
- Science topic