CORRELATION OF POPULATION BIOMARKER IN WASTEWATER

[doctor biom]

Dear,
I am currently investigating Vanyllilmandelic acid as a population biomarker to assess (relative) population. I have stumbled accross a problem since spearman correlation test shows no significant correlation with de facto population (mobile device based). My interpretation would be that VMA levels remain constant in time while population increases and one of the reasons for this could be that area response on the HPLC device was low for VMA (50 for 2000 ng/L and remains below 1000 even at 32000 ng/L). Could the low response be the reason for an underestimation of concentrations? or am i interpreting this wrong?

[tom jupille]

You are interpreting it wrong.

First, you need to make sure of your calibration. Specifically, what are the Limit of Detection ("LOD") and the Limit of Quantification ("LOQ"). You should also identify the linear range, and if the linearity is poor consider using a non-linear calibration.

If your experimental results are well within the linear range (i.e. above the LOQ), I would confirm adequate reproducibility at a "typical" concentration.

THEN you can conclude that VMA concentration does not depend on population and that "low" VMA response is NOT the cause.

[doctor biom]

Dear,

all the cocnentrations are within the linear range. Recovery experiments show satisfactory recovery values accross the linear range with appropriate reproducibility (low RSD). It sucks that no significant correlations can be found since previous studies indicate that VMA correlates quite well with population.

[tom jupille]

No, what sucks is that we are discouraged from publishing "negative" results.

You might want to go back and check those previous studies to see how good their correlation actually was!

[Peter Apps]

As well as the anaytical issues, you need to think carefully about whether mobile phones are a good proxy measure for human numbers.