Order of the labels for quantitation

Dear all,
I have created the following two custom fields:
1) Area_corr_EW (Peak, Real):    Area/SampleWeight
2) Difference (Peak, Real):            U0101.%.(Area_corr_EW)-Ref01.%.(Area_corr_EW)

CF 1 (Area_corr_EW) works fine. For the second CF (Difference) I would expect that the calculated result is identically for both injections. But it isn`t. It works only for the injection labeled with Ref01. The value given in the CF "Difference" of the injection U0101 is the same value as in the CF "Area_corr_EW".

Is it because of the order of the labels in the line "Quantitate" in the sample set?
(The order is U* Ref*)

Thanks in advance!

Answers

  • Empower processes sample sets line by line. So at the label labelled U0101, the Ref01 line hasn't been reached yet. So for the line U0101, the formula will return the U0101 Area_Corr_EW value minus a value which hasn't been calculated yet so U0101 Area_Corr_EW - 0, hence why you got that value.

    When Empower moves on to the next line, it already has a value for U0101 so you get the correct value for the formula. If you don't want a redundant value populated for the U0101 line, then tick the "All or Nothing" option in the custom field editor, this will mean the custom field will only populate at the line where all arguments are present. Or if you wanted to edit the formula, you could change it to EQ(Label,"Ref01)* U0101.%.(Area_corr_EW)-Ref01.%.(Area_corr_EW), which will only populate the correct value for the Ref01 label.

    All or nothing is probably more convenient though.

  • I would say a simpler approach for this would be to simply remove the label reference for the U0101.  Using "Area_corr_EW-Ref01.%.(Area_corr_EW)" would take the current injection's value and subtract the Ref01 result.  It would also be far more robust for any subsequent/potential labels as you only need one formula for any downstream sample label to compare to Ref01.

    If this results in many results where it is not "valid" and you want them excluded from a report, either do something like Empower2019 indicates where you can do an EQ test or simply filter them out on the report so only the valid numbers are reported.