11 Replies Latest reply on May 21, 2018 3:36 PM by ryoboyle

    Overlaid Calibration Curves?

    jtaylor7

      OpenLab CDS Rev. C.01.07.SR1 {113}

       

      Hello again

       

      We're currently experiencing an issue where the number/combination of standard sequence lines selected (1-5) changes what we see in the intelligent report's calibration table. I've attached a few examples:

       

      "Std (all)" is highlighting all five standards

      "Std 1&5" is highlighting either std 1 or std 5 individually

      "Std 2" is highlighting only std 2

      "Std 3" is highlighting only std 3

       

      We've never seen this issue in the past but have been recently working on the templates and think there may be a setting being missed. It's almost like each individual standard is creating its own curve.

        • Re: Overlaid Calibration Curves?
          ryoboyle

          Hello,

           

          There a lot of different possible explanations for this behavior. Is this a report similar to the default Calibration.rdl report?

           

          If you right-click on the Calibration Curve in the Report Layout view and select Properties, check the display format options in the Multi Curves section (below):

           

            • Re: Overlaid Calibration Curves?
              jtaylor7

              We've always had the "overlaid" option selected, but it's never come into play since we usually only see a single curve anyway. I was under the impression the single curve comes from the single calibration table, which is why I'm confused we're seeing multiple curves.

                • Re: Overlaid Calibration Curves?
                  ryoboyle

                  If you are updating the calibration curve in your sequence by designating some injections as calibration samples, then the same calibration curve is not necessarily used for each injection. Also, each compound in your calibration table will have its own curve. With just the screenshots you provided, I can not determine what these different curves are that are being overlaid.

                   

                  Is the calibration curve snippet contained within a group on your report template? If so, if you look at the group properties is the group being repeated?

                   

                   

                  You also mentioned that you only started seeing this after some edits were made to your report template. Do you have any backups made of a known working version of your template that you could try loading?

                    • Re: Overlaid Calibration Curves?
                      jtaylor7

                      Better screencaps attached. They show the entire group (curve and statistical data) both "overlaid" and "separated". I also checked the group repeat and it was set to "=-1". I tried setting it to "none" and there was no difference. Our calibration table identifies a single compound (nitrate) at five levels so what we expect is a single curve consisting of five data points.

                       

                      I was mistaken on the reprocessing bit. I loaded the original sequence and it was occurring prior to any changes being made, and in fact the populated statistical data was way off which probably led to my trying to reprocess it. See "revision 1".

                       

                      As an aside, in "table+curve", which is the calibration table used in "overlaid'' and "separated", why is the curve equation different from that seen on the intelligent report?

                       

                      Thanks again

                        • Re: Overlaid Calibration Curves?
                          ryoboyle

                          Try setting the group to repeat on Compound Name, Injection ID, or Signal Description. This will help us to narrow down where the other calibration curves are coming from.

                            • Re: Overlaid Calibration Curves?
                              jtaylor7

                              Hi Ryoboyle,

                               

                              Apologies for the delayed response. Setting group repeat to Compound Name and Signal Description made no difference. However, setting it to Injection ID resulted in 10 separate curves, presumably due to the 5 standards, 1 control, and 4 samples I had selected. It looks like Standards 1-4 all have different curves, but Standard 5, the Control, and the 4 samples all have the same curves. To double check I sent the latter six to the intelligent report (group repeat set back to =-1) and only ended up with one curve, as expected. See attached.

                               

                              1.PNG= Std 1-3

                              2.PNG= Std 4-5, Control

                              3.PNG= Sample 1-3

                              4.PNG= Sample 4

                                • Re: Overlaid Calibration Curves?
                                  ryoboyle

                                  Thank you for the update.

                                   

                                  That is normal to see the calibration curve change as you are updating it in a sequence. With each calibration standard that you are using the update the calibration curve, you are altering the curve from version of the curve that was used to analyze the previous sample. If you only want to see the final version of the curve that is used on your controls and samples, you could set up a filter on the group to filter out calibration samples:

                                   

                                  The above filter is preventing any data file with a sample type =1 from being reported in the group. As you can see on page 186 of the guide linked below, a sample type =1 is used for calibration samples.

                                   

                                  https://www.agilent.com/cs/library/usermanuals/public/RTEConceptsGuide.pdf

                                    • Re: Overlaid Calibration Curves?
                                      jtaylor7

                                      Hello,

                                       

                                      I responded with another comment prior to seeing your most recent.

                                       

                                      The ten curves above were only seen when group repeat was set to Injection ID. We have group repeat set to =-1, which has always only shown us a single curve. When the sequence was fixed, "=-1" showed a single curve and "Injection ID" showed ten curves, all identical to that seen in "=-1".

                                       

                                      My concern is that I'm still not sure of the cause of the overlaid curves (e.g. the five nitrate data sets all overlapping) or what I ultimately did to fix it.

                                  • Re: Overlaid Calibration Curves?
                                    jtaylor7

                                    In the process of reintegrating a few peaks (via integration off/on and peak reject events, followed by saving to data store) the issue seemed to resolve itself for this specific sequence. There is only one curve for all ten injections now, identical to that seen for the latter six previously. I tried the same thing on revision 1 of the sequence (prior to reprocessing) and it also seemed to fix the issue.

                                     

                                    I then loaded revision 3 (the revision prior to the reintegration performed above) and simply saved to data store without doing any reintegrating. This seemed to fix the issue by itself. I'm not sure why though, considering all of the original integration event changes had already been previously saved to data store.

                          • Re: Overlaid Calibration Curves?
                            jtaylor7

                            I think I may have determined the cause. I noticed on the intelligent report (the most recent one that I believed to be fixed and accurate) that the calibration curve area values were different from the actual peak areas. This was because after the reintegration I forgot to rebuild the calibration table and it still contained the old values.

                             

                            After rebuilding, prior to saving to data store, I checked the intelligent report again. It still contained the old values.

                             

                            I previewed report for a single injection (std 1) to get a save to data store prompt. After saving, I checked the intelligent report again and it had two different curves overlaid.

                             

                            I repeated the above step once again and got the same result.

                             

                            After previewing all of the injections, it went back to being a single curve.

                             

                            It looks like in this case it occurred because I didn't update ALL the standards in data store, so the original issue may have been similar. Does this make sense as a potential cause?

                              • Re: Overlaid Calibration Curves?
                                ryoboyle

                                It is definitely a possibility. Your testing seems pretty thorough and you were able to reproduce the problem.

                                 

                                Once you have confirmed this, it would be appreciated if you could come back and mark your reply as the "Correct Answer" to increase its visibility.