Custom Field Question regarding bracketing standards

I am able to use bracketing standards to calculate our results and set up a custom field to multiply the result by 100.   Now I need to average results for each assay prep (duplicate injections), which I have labeled U0101 and U0102 (and U201 with U202, etc.).
It seems like this would be a result level calculation.  Yet it’s not working.  The system returns a value for the first injection of the pair but not for the second injection and does not average the two values (see below, in which a custom field calc for U101 and U102 would return only the value for U101).
Also, is it possible to put in a sample weight and dilution factor for the standards and have empower calculate the concentration / use the value in calibration?  I’m not able to do this but I may have set this up wrong.I've included some information about my system set-up.


Here’s my sample set:

Best Answer

  • Accepted Answer

    Hi Lisa,

    I forgot that would happen. You have 1 of two choices (and again I'm working this out in my head so I'll need to test it to make sure):

    1. Your report template will be set up to report the individual % Assay values in one table and then ANOTHER table with a data filter to report ONLY report the label U0*02. That way only the correct value (which is always correct for the second injection) will be reported.  The custom calculation can then use the U0?02 label for the % difference and work away from there.

    OR

    2. You use an Enumerated peak custom field which would be something like this...

    ENUM(EQ(Label,"U0102"),EQ(Label,"U0202"),EQ(Label,"U0302"),EQ(Label,"U0402"),EQ(Label,"U0502"),EQ(Label,"U0602"),EQ(Label,"U0702"),EQ(Label,"U0802"),EQ(Label,"U0902"),EQ(Label,"U01002"))

    You can keep going as long as you want depending on how many duplicate samples you're likely to run at the one time.

    Select the USE AS FIELD option in when setting up (see attached) and in the translation table enter the field that you were reporting (let's say it's Average_Assay) followed by (fc), i.e.

    Average_Assay (fc)

    Copy this down for all the rows of the table.

    This way, if I'm correct, Empower will only report the Average_Assay result for those injections with the label ending in 02 and not for the others.

    Note: for youre % difference you won't be able to use the Enumerated custom field.

    Try that and let me know how it goes.

    Kind regards,

    Andre

Answers

  • Hi Lisa:

    I'm not totally clear on what you want to do but if you are just looking for an average, say Average of Amount Percent, it would be easier to keep the labels the same.

    So your sample set would have something like U01 for both assay 1a and assay 1b.  You could then have a custom field like this:

    SAME.%.SAME.AVE(AmountPercent)

    This will give you the average Amount Percent for everything labelled U01, U02, U03, etc.

    For your 2nd question it depends on if you're using the build in Empower custom fields (Dilution and Sample Weight).  If so then Empower will automatically apply

    these values to the calculated amount.  If you want to use your own custom field (something like Amount*Weight) then you could create that field and select it

    for the x value under the components tab of the processing method.

    Hope that helps,

    Chris

  • Hi Chris,

    Thanks for the help!  You were on the right path.  The information you gave was helpful and gave one set of correct numbers but not both.

    Since I am using bracketing standards and am calculating results against two different standards, I should have two sets of averages (one against the first set of bracketing standard and one against the second set).

    I'm not sure how the summarize custom fields in the alter sample window operates-if I need to have it in the sample set twice or just once.

    Any insights are welcome.

    Thanks,

    Lisa

    untitled 1.bmp

    untitled.bmp

  • It's tought to troubleshoot over a forum like this.  Are you generating more than one calibration curve?

    It looks like you have two standards SUSP1 and SEP1 in your original sample set but then your amount table

    shows four amounts.  Is there something missing there?

    Looking at your sample set below it looks like you will definitely have two curves.  You will also have two sets

    of results for each sample.  You should see this when you open the result set and expand the injection/channel.

    In order to get your averages you are going to have to create custom fields that filter your results based

    on the some unique identification (possibly calibration id).  Then you will be able to average each set of results.

    Hope that helps.  It's tough to see w/o seeing the data.  You can always call Waters and they are pretty good

    about helping with custom fields if you send them your data (and are willing to pay ;-)).

  • I strongly recommend getting Waters Technical support to help you here.

    Summarize custom fields cannot distinguish the results for a particular unknown from two calibration steps. It only uses Label and " in this result set" so will get a %RSD that includes all matching results.

    We have recently looked at possible including a filter in the custom field comparing Result IDs against Max Result IDs to do something similar. So you would search for when the result id NOT EQUAL max ID to get the USP results and when it DOES EQUAL Max ID for the EP results.

    Me?  I would use the % RSD in a Report, however you still need to split the two sets of results...  A long time ago I did this but at that time you could only filter the table using processing method, so I made two (identical) processing methods, ie Carbazine EP PM and Carbazine USP PM, used these two to calibrate and quantitate the appropriate Stds and unknowns, Then filtered my results in the table by processing method. This sounds complicated, but is much simpler than trying to separate by result IDs and using Summarize fields. (NB you might be able to combine these ideas, use different PMs and try a filter in a custom field as well..)

    H

  • Just a note.. A result level field would say, average all areas in ONE RESULT..  not across results normally.

    Heather

  • Hi Lisa,

    We use the Mean Function available in the Component Summary Table in Reports to mean results from sample replicates.

    untitled 1.bmp

    If we have more than one sample in a standard bracket then we use one Calibrate and Quantitate set for each sample

    <td valign="top" width="...

    5

    S0201

    Inject Standards

    Std B

    Don't Process or Report

    6

    S0301

    Inject Standards

    Std B

    Don't Process or Report

    7

    S0401

    Inject Standards

    Std B

    Don't Process or Report

    8

    S0501

    Inject Standards

    Std B

    Don't Process or Report

    9

    U0101

    Inject Samples

    1713-11-001H, 25°C/60%RH 6M - A

    Don't Process or Report

    10

    U0102

    Inject Samples

    1713-11-001H, 25°C/60%RH 6M - B

    Don't Process or Report

    11

    U0201

    Inject Samples

    1713-11-001H, 40°C/75%RH 6M - A

    Don't Process or Report

    12

    U0202

    Inject Samples

    1713-11-001H, 40°C/75%RH 6M - B

    Don't Process or Report

    13

    S0601

    Inject Standards

    Bracketing Std B

    Don't Process or Report

    Clear Calibration

    Normal

  • Your Q:Also, is it possible to put in a sample weight and dilution factor for the standards and have empower calculate the concentration / use the value in calibration?  I’m not able to do this but I may have set this up wrong.

    Answer:Hi lisa i am giving reference for your post.

    Assay= Amount*Dilution Factor*Potency

    Here Dilution factor gives your factor.

    Example: Your standard dilutions STD1(100ml),STD2(10nl),STD3(100ml)

                    Your sample dilutions  SPL1(100ml),SPL2(10ml),SPL3(100ml)

               

                     So your dilution factor = (STD2*SPL1*SPL3)/(STD1*STD3*SPL2)

                   

                      By using above you can get the dil factor.

     

    Regards

    SATYA

  • Satya-

    I think I understand what you are trying to get at: Essentially you want to use full precision by entering in the mass for your API in the amounts field in your sample set and then use other fields to accurately get your calibration curve rather than entering a concentration in Empower's amount field?

    Empower will automatically take the entered amount values, multiply it by the sample weight field, and divide by the dilution field that have been entered in Empower for calibration purposes (samples are back-calculated opposite of that...multiplied by dilution field and divided by sample weight field).  I've been utilizing this feature for years to capitalize on full precision calculations in Empower to quantify my results.  The trick that I use is that purity is entered as a ratio (i.e. a potency of 98.5% is entered as 0.985) in the sample weight field.  The dilution field is entered as the inverse as what people classically approach dilution as normally as Empower will divide by that number, not multiply.

    I'm not sure I understand your comments about your specific dilution scheme...

    My example:

    Weigh X mg API (98.5% Potency) into 100 mL (Std 1)

    Transfer 1 mL Std 1 into 50 mL (Std 2)

    -If using Std 2 to calibrate, the value you'd enter in Empower for your Dilution field is (100/1*50=5000 for mg/mL concentration, if you want it in micrograms, include unit conversion: 5000/1000=5 for ug/mL).  Potency would be added as 0.985 in the Sample Weight field in Empower. Enter your API mass into your amounts table and you should be set.

    Empower will then utilize these values to calculate the calibration curve.  This will work whether you are using brackets or a full linearity using multiple standards across a range (obviously each standard in that scenario would utilize a different dilution value when they are at different final concentrations).

    Mike

  • Hi Lisa,

    It seems that this tread is getting a lot of attention today and you may have already got your answer. In case you haven't here are my 2 cents worth.

    Firstly, I'll answer the second question - can you put Std weight and dilution into the sample set method.

    Answer - yes you can. Here's what I do.

    FOR STANDARDS, I put the weight in the SampleWeight field. I put the dilution FACTOR in the dilution field. This is calculated by the following calculaton:

    dilution factor = product of the volumetric flask volumes / product of the pipetted volumes

    e.g. if I weigh 5.0120 mg of standard material into a 100 mL flask and then take 2 mL from that flask and dilute to 25 mL the following is put in the relevant fields...

    SampleWeight = 50.0120

    Dilution = 1250  = (100 x 25)/2

    But what about the purity? Well that's easy. I put the purity of the standard in the component editor (Value) field as percentage divided by 100 - i.e. 98.5% = 0.985. This is what I put in the component editor for the standards.

    Empower calculates the amount for standards as the following:

    (SampleWeight x Value in component editor) / Dilution

    As for your custom field - a PEAK field is what you need. The way I view the difference between peak and result fields is like this...

    Peak fields pertain to the individual peaks in a chromatogram (or between different chromatograms).

    Result fields look at ALL the peaks in ONE chromatogram (e.g.  the SUM of all peak areas IN THAT CHROMATOGRAM or the AVEerage of the peak areas WITHIN THAT CHROMATOGRAM.

    For your averaging problem - here's my solution which does require an inter sample summary calculation (reason being it's hard to get a labelling method that will work with wildcards but not take extra injections into the calculation - it's not impossible, but I don't have time right now to think of it)....

    Here's the calculation

    (U0?01.%.SAME(AmountPercent)+U0?02.%.SAME(AmountPercent))/2

    And you wouldn't need the summarize custom fields line. Make sure All or Nothing is ticked and you SHOULD get the result in the U0102 injection (I haven't tested this but it should work).

    Hope this helps you or anyone else.

    Andre

  • Hi Andre,

    Thanks for your help!  The calculation information on purity helped quite a bit in simplfying our calculations.

    The equation you provided is closer to what I need and your explanation on the peak vs result was helpful.  I ran smack into the label issue with intersummary calculations, so it's refreshing to find a way around it.  It's still picking up a few extra injections (see attached, highlighted in yellow).

    Any thoughts?  I suppose I could set up a custom field for each label and call out specific labels (ie U0101 and U0102) and would have three separate custom fields.

    Thanks,

    Lisa

  • Lisa,

    Do you need the Average assay to be in a field in the database?  Or would you be content to use reporting features to calculate the average results (as suggested by Tim)?

    You can Group the data table by Label or some other field, so that in one report you can get three separate tables for prep1, prep2, and prep3.  I would agree with Chris that making the label for prep1 and prep2 the same would greatly simplify things.  There is no rule that says the Labels need to be unique.  Use them as a tool more easily process or report the injections you want.  Unless you would ever want to process one and not the other, having the label be different doesn't really help anything. 

    Hope this helps,

    Kevin

  • Hi Kevin,

    Thanks for your input.

    The average assay calculation would need to be in the database as I need to have the database do a further calculation showing the % difference for two results.  For example, for assay 1 calculated against the USP standards, 1a is 99.0 % and 1b is 100.0 %.  The average would be 99.5 % and the percent difference would be 1.0 %.  If I used the summary report feature and kept the labels the same (U01, U02, U03), I wouldn’t be able have it show the % difference.

    --Lisa

  • Hi Chris,

    Chris, thanks for all your help.   I have two calibration curves for one set of samples and do have two sets of results.  The two standards are USP1, USP 2 and EP1, EP2 and the four amounts in the amount table pertain to those standards.  I may have to consult Empower. 

    How do you filter results based on the calibration ID using a custom field?  I'll play around with it more.  Let me know if it's helpful to see more data.

    Thanks!

  • Hi Heather,

    Thanks for your input.  I've contacted Empower with my questions and am waiting for a response.  In the meantime, could you tell me more about how you set up the custom field search filter? 

  • Hi Andre,

    Method 1 works great!  It works for both average assay and amount percent difference.  I haven't tried method 2, but maybe I'll play around with that.

    Thanks so much for your help!

    --Lisa

  • I'm glad I could help.

    I only stumbled onto this forum a few days ago (whatever day I first replied) and your thread was the first I read. I'm just sorry I wasn't able to access this forum in July - it would have saved you 6 months.

  • Hi,

    The method you've outline so far works great for simple assay calculations.  I ran into a situation where I needed to add the areas of two peaks (group them) to perform calculations.  I hoped the custom fields would act the same way; however, the custom field calculations did not work.  In the grouped line in processing method I selected "x" under the Weighting column.

    Do I need to set up alternate custom fields for this situation or is the processing method not set up correctly?

    Thanks for your help!

  • For summing two peaks, you can very easily utilize a named group within the processing method.  I've had to do this numerous times.  Under the named group tab for the processing method, click in the table to create a new line item and name it (sumXY or whatever you want).  That name will then appear in the box on the left where you can then drag the peak name(s) for the peaks you want summed to the newly created field.  This can be used in a way that amount values can either be directly entered for the named component or where this new component can utilize a completely separate peak for quantitation so make your selections carefully under the "source of calibration" column.

    The named component can be utilized in custom fields just like any other peak name, if desired.  Just be careful to not utilize any punctuation or custom field codes that would confuse the custom fields (e.g. mathematical symbols like + or -, parentheses, etc.) within the name of the named group.

  • Hi there,

    Thanks for your response.  I followed your reccomendations to copy the group peak from the PM to the Component editor, and still empower cannot seem to utilize the group peak properly to carryout calculations.  Do you have any advice?  I updated my query with snapshots of what I've set-up and the results.

    Thanks,

    Lisa

  • Hi Lisa,

    I think all you need to do is instead of having Sum peaks, curve or Sum peaks for quantitation in the Named groups tab of the Processing Method change it to User Entered, Curve or Sum peaks for quantitation.  Then, like you tried, enter the concentraiton (or purity value if you've utilsed the sampleweight and dilution fields for the standards) in the Component Editor for the GROUP PEAK. A calibration curve will be drawn for the grouped peak and the grouped peak in samples will be quantified off that calibration curve AND the custom field should work.

    I tried it myself and it worked fine for me.

    SO I think you're practically there - just change the Source of Calibration X Value in the Named Groups tab of the PM to be User Entered and I think you should be good to go without changing anything else.

    Let me know how that works.

    Andre

  • Hi Andre,

    Your response was helpful as it led me to the problem.  I could see the calibration curve for the grouped peak and the calculated amount for the samples.  However, the custom field result still didn't work, which meant there was something set up incorrectly in the custom field.  The peak type in the first custom field had "founds only" selected, which meant that the custom field couldn't find the grouped peak,  Once the peak type was changed to "all" the custom field calculations worked for the grouped peak. 

    Thanks again!

    regards,

    Lisa

Sign In or Register to comment.