9 Replies Latest reply on Oct 30, 2018 2:25 PM by ryoboyle

    Calibration in OpenLab DA

    julespickles

      Hi,

      I'm currently setting up our first system using OpenLab CDS 2.2 including the Data Analysis module. I'm struggling to understand how the calibrations work, hope someone can help with a couple of questions:

      1) I've set up an external standard processing method for my validation study, which seems to have worked ok, in that the samples have all been integrated and quantified, however I need to exclude one of the standards as the %relative residual is outside my acceptance criteria. When I exclude the point, it only applies to one sample, and remains in the calibration for all others. I see from the Help and Learning tool that that is by design. How can I get the exclusion to apply universally?

       

      2) When I generated a report for my samples, there were 3 versions of the calibration line displayed (I assume due to my fiddling about trying to solve 1) above!), and the calibration table did not display the correct %relative residual information, despite me excluding the inaccurate point for all samples individually. I only want to display the most recent line and table. How can I configure that?

       

      Thanks in advance for any help!

        • Re: Calibration in OpenLab DA
          mr.white

          Recording Information 

          She Show you the Basics of filtering within RTE.

          • Re: Calibration in OpenLab DA
            julespickles

            Can anyone help with how I can update my standard exclusion to the method version of the calibration? I can't even seem to view it in data analysis, but that's what is appearing in my report, how can I amend it?

            • Re: Calibration in OpenLab DA
              ryoboyle

              Hello,

               

              I added tags to this post to increase visibility.

               

              1) You are correct, when you disable a calibration point in OLCDS DA that disabled point only applies that the currently loaded data file. I have pasted the relevant section of the Help and Learning which you referenced in your original post below that others can reference it quickly:

               

              If you wish to remove that calibration point from your cal curve for all of your sample injections linked to this processing method, then you will need to reprocess your result set. Before reprocessing you will need to make sure that your calibration standards are identified as such in your injection list, but you want to change the sample type for the data file associated with the cal point you wish to remove from Cal. Std. to Sample. This way, this data file will not be used to update your calibration curve when you Reprocess. You also want to make sure that you set the Run Type of your first Cal. Std. line in your Injection List to "Clear all calibration". This will remove your current calibration curves from you method and build new calibration curves from your Cal. Std. samples the next time you Reprocess.

               

              Let me know if you have any questions.

               

              2) Are you able to post a screenshot/picture of this problem? Also, what type of report template are you using: Single Injection or Sequence Summary?

              • Re: Calibration in OpenLab DA
                julespickles

                Hi,

                Thanks for your response, that's really helpful. I've tried what you suggested eg setting the standard I wanted to exclude as a sample, and reprocessing with clear all calibration for the first standard in the line. I have a couple of other questions now though!

                 

                Before I reprocessed, all standards were visible in the calibration lines for each sample.Now, each standard only has the calibration points that were injected before it included. This means that the first sample (where it was set to clear all) isn't quantified at all, and the second one has a perfect regression as there are only 2 points in the line. This leads me to believe that the standard concentrations are being calculated against these partial lines, is that correct?

                 

                I would expect the standard concentrations to be calculated against the full standard line saved in the method, not in the partial ones created during processing. Is it possible to set it up this way? I need to be able to easily view the standard accuracies against the line that will be used to quantify my samples to ensure they are meeting acceptance criteria. I don't want to have to recalculate the standard concentrations in Excel using the calibration equation. Perhaps I'm misunderstanding, as I'm used to working in MassLynx/TargetLynx where there is only one calibration line for all injections?

                  • Re: Calibration in OpenLab DA
                    ryoboyle

                    Hello,

                     

                    You are correct. In a standard calibration sequence with no bracketing, the results for an injection's data file are generated immediately after acquisition/processing with the version of the processing method which is available at that point in time. There are two possible ways that I know of which would allow you to generate results with the "full standard line" as you referred to it.

                     

                    The first would be to specify a bracketing mode other than 'None' in your Injection List or Sequence Template. There are different bracketing modes available to choose from (descriptions available in the OpenLab Help & Learning), but in short bracketing will wait until multiple calibration samples are run before updating the calibration table and generating results.

                     

                    The second method I can think of is to change the Sample Type of your calibration standards from Cal. Std. to sample after your calibration curves have been updated. Then the next time you reprocess the data the software will use the Processing method with the "full standard line" to process all of the data files and generate results.

                     

                    Hopefully this helps.

                      • Re: Calibration in OpenLab DA
                        julespickles

                        Hi,

                        Thanks for the advice, I've been trying it out, but with limited success sadly!

                         

                        I still don't understand this which I posted above: "Before I reprocessed, all standards were visible in the calibration lines for each sample.Now, each standard only has the calibration points that were injected before it included. This means that the first sample (where it was set to clear all) isn't quantified at all, and the second one has a perfect regression as there are only 2 points in the line etc etc." Using different bracketing techniques didn't change this.

                         

                        I've done the following sequence of actions:

                        1) Load and process the data (all standards showed all calibration points in their calibration lines)

                        2) Set the outlier standard as a sample, and the first standard in the line to Clear all calibration and reprocessed

                        3) This gave an acceptable calibration (in terms of %residuals) for the top standard/unknown samples where all the standards were included, but now the lower standards don't have all the points in.

                        4) Change all standards to samples and reprocess - this appears to quantify the samples correctly, but I can't get the calibration information to display properly in my report, as there are now no standard samples in the run, and there's no history % residual field in the table, only history amount and history response. %residual is an important acceptance criteria for us, so I don't want to have to calculate it manually.

                         

                        As far as I'm aware what I'm trying to do is not that unusual. Is there really no way to view/update the method version of the calibration within the DA window (as I believe was the case in Chemstation?), and set all samples to process against that version? For data integrity purposes we're not that happy with the idea of individual calibrations for each sample being able to be modified (either intentionally/accidentally!).

                         

                         

                         

                          • Re: Calibration in OpenLab DA
                            camelia

                            in the Injection results window select Bracketing mode then using the drop down arrow select Custom, add an opening bracket [ to the first Cal standard and a closing bracket ] to the last cal standard. Use Clear all calibration on Cal standard 1 as before.  Exclude any standards if permitted by applying the sample run type instead of standard as before. When the data is reprocessed the calibration curve equation should now remain the same as you move through each injection as it is using all of the calibration points. Hope that helps.

                              • Re: Calibration in OpenLab DA
                                julespickles

                                Thank you for the suggestion. I've tried it, but I get the same result as before: when I initially process the standards, all are included in the line (start bracket on std 1, end on std 10, clear all calibration set on std 1). When I set the excluded standard to sample and reprocess, each standard has a different calibration again. It seems to be a problem with having a sample in the middle of the standard bracket. I tried setting it with two brackets 1-7 and 9-10 (8 is excluded), but then stds 1-7 were missing 9-10 from their calibration as they were in their own bracket.

                                  • Re: Calibration in OpenLab DA
                                    ryoboyle

                                    Sorry for the delay. Having samples within a bracket is a normal workflow (see screenshot of an example Non overlap bracketing injection list), as otherwise there would be no difference between bracketing and not bracketing your calibration:

                                     

                                     

                                    If bracketing is not working for you as designed, then please reach out to your local Agilent support for help: Contact Us | Agilent