22 Replies Latest reply on Feb 27, 2018 6:52 AM by ryoboyle

    Multi-level calibration

    mjebrane

      Hi,

      I'm using EZChom elite software, and trying to calibrate a method (multi-level calibration). My problem is that when I put graphically all the levels (2 replicates  for the first compound, and when I put the levels for the second compound, and when I review calibration, the replicate of the compound 2 is found in the curve of the compound 1. I have tried to change different parameters in the peak table, but without success.

      Please if you have any solution help me.

      Regards,

      MJ

        • Re: Multi-level calibration
          ryoboyle

          Hello MJ,

           

          I apologize for the delay in response. If you still need any help with this EZChrom Elite calibration issue, would you mind letting me know what revision of the software it is that you are using?

           

          Best Regards,

          Ryan

            • Re: Multi-level calibration
              mjebrane

              Hello Ryan,

               

              Thank you for your email. Yes I still need a help regarding multi-calibration problems. I’m using EZChrom Elite version 3.3.2 SP2 (build 3.3.2 1037).

               

              Best regards,

              Mohamed

                • Re: Multi-level calibration
                  ryoboyle

                  Hello Mohamed,

                   

                  Thank you for the info!

                   

                  My recommendation would be to first, if you are satisfied with how your peaks are being integrated, identify your compounds in your peaks table. My preferred method of doing this is to right-click on the chromatogram and select Graphical Programming>Define Peaks and follow the prompts for selecting a time window. This will pull RT information of all integrated peaks in your selected window into the peaks table where you can then name the compounds, select your curve fit options, and enter the amounts for your calibration levels.

                   

                  Once all of your peaks of interest are defined in the Peaks/Groups table, now you need to actually build your calibration curve(s). I recommend doing this through your sequence table. In the sequence table you can define the run type as a calibration run and select to average replicates in the 'Run Type' column. You would also identify what level each run in the sequence table corresponds to in the 'Level' column. Once these fields are filled in, then you need to process the sequence to update your calibration table by selecting 'Sequence'>'Process'.

                   

                  The act of processing the sequence will build the calibration curves for your compounds as you have specified in the Peaks/Groups table and your Sequence table. Please let me know if that does not work for you or if you have any questions.

                   

                  Best Regards, 

                  Ryan

                    • Re: Multi-level calibration
                      ryoboyle

                      mjebrane, I just wanted to follow-up to see if my last suggestion helped or if there were any other questions.

                        • Re: Multi-level calibration
                          mjebrane

                          Hello Ryan,

                          Unfortunately I run to the same problem, even following strictly your procedure, all compounds were mixed up in the calibration curves (see image below: all other compounds were mixed with Mannose calibration curve for all levels). I have double checked in methods properties that consecutive replicates for same level should not be averaged but I got same results. Please if you have any explanation send it to me.

                          Many thanks.

                          Mohamed

                            • Re: Multi-level calibration
                              ryoboyle

                              Hello Mohamed,

                               

                              Have you previously tried to build a calibration on this same method? If so, are you choosing to Clear All Calibration in your sequence table so that you are starting off your new calibration from a blank slate?

                               

                              As long as you clear the calibration at the beginning of your sequence (if there is already old calibration data saved to the method) and then specify to Average Replicates in the sequence table then you should be fine. Having the method properties set to automatically default to averaging calibration replicates is good to have selected as a safety net.

                               

                              In the screen capture you posted in your last reply, the responses appear to be varying quite a lot. Have you double-checked to make sure that you are specifying the level correctly for each line in your sequence?

                                • Re: Multi-level calibration
                                  mjebrane

                                  Hello Rayn,

                                   

                                  No, this is a blank method, and I always clear all previous calibrations. In the posted screen capture, there is all replicates from all compounds for each level, which is not normal, I should get only calibration level for the specified compound.  I have checked everything and I’m sure that each level is specified for each compound.

                                  Just for info, I have been doing this for  long time and always it works pretty well, just recently I run to this problem, which I can’t figure out.

                                  Best regards,

                                  Mohamed

                                • Re: Multi-level calibration
                                  danots

                                  clear calibration settings.png

                                  I would expect you are already familiar with these settings, but I am pasting a picture just in case.   The calibration above appears to have several older calibrations.  On the right side of the first line of the sequence in the Run type box, click the blue arrow and check the Clear calibration box to remove all old calibration curves.  Check Average replicates if you want to average the replicates.    When done, there should only be one point for each calibration level.

                                    • Re: Multi-level calibration
                                      mjebrane

                                      Hello danots,

                                      Thank you very much for sharing your thoughts.

                                      if I check the Clear calibration box, all the calibrations disappeared, and if I calibrate again by checking only the average replicate box I run to similar problem.

                                      Best regards,

                                      Mohamed

                                        • Re: Multi-level calibration
                                          danots

                                          I was recommending you check both boxes at the same time and only process once.    I cannot read the sequence information in the posted picture. Can you post a copy of your method so I can look at the calibration information?   Can you print your picture into a PDF file so I can get a better look at the sequence info?

                                          • Re: Multi-level calibration
                                            danots

                                            Also if it is not clear, average replicates should be check for each line, not just once at the beginning.  You also need to check begin calibration on the first line and end calibration on the last line.

                                              • Re: Multi-level calibration
                                                mjebrane

                                                Hi danots,

                                                Yes I have checked both boxes at the same time (Clear calibration and average replicates), but all calibrations disappear.

                                                I use to check average replicates box for each Line and check begin calibration on the first line and end calibration on the last line.

                                                I have build a new method and double checked all the parameters but unfortunately I run to the same problem.  here is the picture of part of the method report and calibration curve for one compound and after starting calibration for next compound.

                                                Thank you.

                                                  • Re: Multi-level calibration
                                                    danots

                                                    The curve in the attached method reports looks the way I would expect it to look.  Is there still something wrong with the calibration.  Both reports look correct to me. 

                                                      • Re: Multi-level calibration
                                                        mjebrane

                                                        if you look at  Arabinose Galactose Method report you will see that at the end of the first column (level 1) it includes also area of Galactose and the same in the calibration curve (3 replicates instead of 2 replicates). this is what is wrong actually.

                                                          • Re: Multi-level calibration
                                                            danots

                                                            This means the Galactose std 0,3 mg.ml new2017-07-01 06:48:35 data file is listed as a calibration file and is being added as a replicate.   Since I cannot see the actual calibration sequence and read all the settings it is hard to point to what is wrong.

                                                             

                                                            Here is what I am expecting.  Your sequence has a list of 10 data files.  5 files are levels 1-5 and the second 5 files are listed as levels 1-5 and area being averaged with the first 5 levels.

                                                            The first line of the sequence clears the calibration curve and has the begin calibration command.  All lines after this are set to average calibration and the last line has the end calibration command.  The level box has the numbers 1-5 for each set of data files.  This 1-5 numbering is duplicated twice.  Any non calibration files are listed in the sequence after the end of the calibration command.

                                                             

                                                            How many data files do you have in your sequence?

                                                             

                                                            Thank you,

                                                            Daniel Richmond

                                                              • Re: Multi-level calibration
                                                                mjebrane

                                                                Actually if do calibration using sequence table, I get all compounds replicates mixed up, so I build a calibration one compound by one. In the case I sent to you, I have a calibration curve for arabinose with two replicates for each level (5 levels), but when I start same process for the next compound (galactose), the data of galactose is automatically added as replicate in arabinose calibration table/curve for all levels and for all other compounds (in the method I have sent to you, I have only Galactose std 0,3 mg.ml new2017-0701 data which was automatically added as replicate to Arabinose calibration curve, I haven't finish calibration for galactose and other compounds because there is no point, I will come to same problem: all compounds will be mixed up).

                                                                regarding your second part of your comment, yes it exactly what you wrote: my sequence has 10 Data files for each compound, 5 files are levels 1-5 and the second files are listed as levels 1-5 and are being averaged with the 1st 5 levels. I have 5 compounds, so in total 5 i have 50 files in the sequense table.

                                                                i will send you tomorrow my sequense table that you can understand better.

                                                                thank you for your help.

                                                                  • Re: Multi-level calibration
                                                                    danots

                                                                    I do not know of a process to correctly process one compound at a time using the same set of data files to be reprocessed for a new compound each time.   When you process the calibration it should be for all compounds at the same time.  You should only process the sequence once.

                                                                     

                                                                    Thank you,

                                                                    Daniel Richmond

                                          • Re: Multi-level calibration
                                            ryoboyle

                                            I went ahead and marked this as "Assumed Answered" since there has not been any activity in a while.

                                             

                                            Please let us know if this is still unresolved and we would be happy to continue helping you.

                                              • Re: Multi-level calibration
                                                mjebrane

                                                Hi,

                                                Unfortunately, it is not yet solved, i have been in touch with the service engineer but couldnt figure out the problem.

                                                 

                                                Best regards,

                                                Jebrane

                                                  • Re: Multi-level calibration
                                                    ryoboyle

                                                    Hi Jebrane,

                                                     

                                                    I am sorry to hear that. The behavior that you are describing is definitely not how the software should operate. As long as you follow the workflows that danots and myself outlined in our previous comments then there should be no difficulty in building a calibration curve.

                                                     

                                                    I encourage you to continue working with your local service engineer as they should be able to witness this behavior first hand either onsite or via a remote webex session. If you want to do any more troubleshooting through the Community, then we will need a copy of your result set including your data, method, and sequence.

                                                     

                                                    Also, what operating system is this installed on? Windows 7? 32bit or 64bit?