# System Suit requirement using Internal Standard

I run a method with two actives, Active A is at 220nm and Active B is at 235nm. I also spike Internal Standard into my standards and samples. I have two different derived channels for the 2 wavelengths and hence two processing methods. I have taken over this method from a co-worker who has left the job and he always used an Excel spreadsheet for the various calculations and I'm wondering can Empower just do this much quicker..

One of my system suit specs is as follows:

Two injections of Working Standard- Working Standard 1 and Working Standard 2 (both prepped separately but with the same concs of actives and IS in each). The relative difference in responses over standard 1 and 2 needs to be 2% or less. The Excel sheet is set up to display the Area of the Active peak, the Area of the Internal Std peak and the concentration of the Internal Std (in mcg/mL). The calculation is: Area of Active A/(Area of Internal Standard*Concentration of Internal Std). Now this seems backwards to me, I thought Internal Standard Response was (Area of Active/Area of Internal Std)*Concentration of Internal Std? The first calculation gives a very different number to the second one for example Area of Active is 87665, area of IS is 56009 and conc of IS is 159.007 mcg/mL. The Excel sheet gives a value of 0.00984 for response whereas surely the correct value is 248.8769?

If I did this by Empower, would I be right in selecting the IS component as the "Internal Std" peak against my active in the processing method, then putting the conc value of the IS in the comp editor for the standard injections? Would Empower then display response value in the result set for these injections? I'm only interested in these specific two standard injections for the time being, not samples.

Can anyone shed any light please?

• I have a method where it is similar...basically calculating the ratio of bracketing standards throughout the analysis (so standard injections with an IS, comparing the responses).  From what you are saying, you do seem to be on the right path both Empower-wise and what I would just consider correct math.  Enter the applicable values in the component editor within the sample set.  Then specify the internal standard by name in the processing method's components tab.

When using an IS, Empower will calculate a response factor (field name=RF).  The gap I see with what you've detailed though is that the formula is missing the ActiveA concentration from the component editor for this.  Empower would calculate as follows for the RF field: AmtIS/AreaIS*AreaActiveA/AmtActiveA (where Amt is the amount entered in the component editor).  You would really just then need an intersample calculation comparing the two RF results.  Since the two actives are in different channels, then for sure you should be able to have one formula that calculates both using the "SAME" in the intersample calc for the channel specification.
• It seems like your follow up has 2 parts...how do I normally calculate assay? And which is the best practice for brackets, calibration or QC drift checks?

For the few methods that I have in my lab that utilize an internal standard, we do exactly what I stated above to generate the RF values which, in turn, generate an amount.  All of our assay tests (internal std or not) basically utilize Empower's components and such to generate the amount field in a fully validated, non-custom field, way whenever possible.  I imagine that this is a very common approach as it brings about so much efficiency when it is all done with fairly minimal analyst training.

For brackets, I don't think I want to speculate on how everyone else does it, but we've typically written our methods to utilize brackets as the calibration source.  Which you would prefer probably depends on what sort of perceived or allowable drift you have in the method.  I like the approach of using the brackets for calibration as it can help compensate for any system drift throughout long analyses.

For the %Assay, your overall outline is exactly how I'd do it...unless the nominal values are always constant, I'd have sample fields for the analyst to enter the "ActiveX Nominal" and "ActiveY Nominal" in the sample set, then a custom field for the % (i.e. %X=Amount/ActiveX Nominal*100), then sort the data accordingly on the report.