Is it possible to process multiple sample sets and get one result?

<p>We have a method that requires the following:</p><p></p><p>Run 10 samples, if the criteria for these 10 samples fails, run 20 more.  We then have to calculate an average using all 30 samples.</p><p></p><p>The problem I see (I haven't tested this yet) is that we will have 2 separate sample sets that we will need to do calculations on.</p><p>Has anyone done this and been able to get results using both sample sets?  If so how do you go about processing and doing the calculations?</p><p></p><p>Thanks,</p><p>Chris</p>

Answers

  • Hello,

    if you have Empower version 3, it is possible to create Process Only Sample Set in your project.

    After that you can process Process Only sample set data together in one set and you will obtain results for all injection included into inter-sample calculation.

    It is useful for some MVM project testing, dissolution calculation, when you collect only 1/2 of your Sample Set and you need to add the second part measured separately ... etc.

    Jiri

  • Aside from the upgrade, the only thing that you might be able to do is add the additional 20 samples to the sequence, select those additional samples and choose "run selected lines only" when you go to start the run. From there, you will likely have data in two sample sets having the same name, but started at different times. If you select these two sample sets and process as you normally would they might combine into one result set.

  • The devil is in the detail!


    Something that might work would be to keep the samples in two sample sets, but to give each set of 30 samples the same label, making sure that each set of 30 samples had a different label.

    You could then create a custom field set to "Result Set First", with the calculation SAME.<injection replicate number or % for all injection replicates>.<channel/channles or SAME>.<summary calculation, e.g. AVE, MAX, or anything more complex>


    There are a few "gotchas" in there though - for example, if you re-process the set, you'll want to assign new labels first - otherwise both the original and revised values could be included in the result. Reprocessing individual results could cause problems too. You might struggle to quantitate the samples in the individual runs when they all have the same label, but that'll depend on your run structure. You might need to use custom fields to identify signed or accepted results. Depending on the calculations you use for your criteria, you might be able to use the EQ/NEQ/LT/GT functions to work around a few of these problems though.


    Alternatively, ShaunWat78's idea of creating a 30-sample run would work really well if you trigger an error/abort after the first 10 samples. Perhaps add a dummy sample to the run with a different (or toxic) method set after the first 10? You can then resume the sample set method later if the first ten samples pass, and by selecting the right options on restart you can finish up with one sample set and all 30 samples...

    If your acceptance criteria calculations can be done using Empower report method table summary calculations, you can select the 30 sample results individually and report them. You probably already knew that, but thought I'd mention it just in case.


    Empower 3 is a good solution though!


    If things get desperate, there might be another way...


    Hope this helps!

  • Hi Chris,

    I met the same issue (CU). I use Empower2 FR5. The bypass solution we have is to use LIMS for these types of calculation. So outside Empower.

    From my point of view, Empower gives us results and LIMS gives us the conformity. So I'm not completely upset without an Empower solution.

    regards,

    Brice.