Making metabolite samples with different volumes comparable prior to paired T-test

#1
Hi,

I have metabolite measurements from tumor patients comparing samples from the tumor tissue (3 replicates per patient) against nearby unaffacted control tissue (again, 3 replicates per patient). Since the tumor and control data for each patient is paired and the overall measurements are approximately normally distributed, I would like to use a paired t-test to compare whether a metabolite abundance differs significantly between tumors and controls across all patients.
However, the samples for different patients have slightly different volumes, so I need to make them comparable first. Does a simple standardization (mean centering and scaling to variance 1 for each patient, tumor and control samples together) do the job, or do I only need to subtract the maximum and divide by the difference between maximum and minimum for each patient?
Or can you even recommend a better statistic to compare the differences between tumor and control samples than the paired t-test?

Many thanks in advance.
 

chiro

New Member
#2
Re: Making metabolite samples with different volumes comparable prior to paired T-tes

Hey garfield42.

How is the relation between the concentration of the thing you are measuring and the volume? Is it linear or non-linear?

If its linear then what you can do is to scale it by a constant and if its say proportional the the change in volume you can scale it by V1/V2 where those are the two volumes involved.