GATK Tutorial#11682 reproduce different results

[yuweibao15]

Brief issue description:
When following the tutorial https://gatk.broadinstitute.org/hc/en-us/articles/360035531092--How-to-part-I-Sensitively-detect-copy-ratio-alterations-and-allelic-segments, the #4 Plot standardized and denoised copy ratios with PlotDenoisedCopyRatios have different results than the tutorial. Through the control vectors test, it seems that the samples that are used in step #2 to generate CNV PON used in the tutorial are different from the files stored in the tutorial.
Results:
Following steps 1 to 4, the resulting plots


The results have values However, the values in the tutorial are 0.134 and 0.125.
Tests
Using the files provided in the tutorial and script generated cnvponC.pon.hdf5, which seems to lead to this inconsistency result.
Using:
gatk --java-options "-Xmx6500m" CreateReadCountPanelOfNormals
-I HG00133.alt_bwamem_GRCh38DH.20150826.GBR.exome.counts.hdf5
-I HG00733.alt_bwamem_GRCh38DH.20150826.PUR.exome.counts.hdf5
-I NA19654.alt_bwamem_GRCh38DH.20150826.MXL.exome.counts.hdf5
--minimum-interval-median-percentile 5.0
-O sandbox/cnvponC.pon.hdf5
Files
The script used to generate this result are attached.
gatk_tutorial11682_issue.zip

Please help me understand this difference in reproducing the tutorial result. It will be extremely helpful for me to use the pipelines on our lab-generated data. Thank you very much!

[yuweibao15]

With further inspections, step #3 DenoiseReadCounts produced different values inside hcc1143_T_clean.standardizedCR.tsv and hcc1143_T_clean.denoisedCR.tsv compared to the results stored in the tutorial. This could lead to a possible cause of the file hcc1143_T_clean.counts.hdf5 provided in the tutorial.

[droazen]

Hi @yuweibao15 , could you please post this question on the GATK forum (https://gatk.broadinstitute.org/hc/en-us/community/topics)? We'll be able to assist you there.