PlotDenoisedCopyRatios lead to empty result for non-human results

Follow

Yuwei Bao

• June 23, 2024 21:07
• Edited

REQUIRED for all errors and issues:
a) GATK version used: docker image spacecade7/tutorial_11682_11683:gatk4.0.1.1

b) Exact command used:

docker run -v WORKDIR/./:WORKDIR/./ spacecade7/tutorial_11682_11683:gatk4.0.1.1 gatk PlotDenoisedCopyRatios --standardized-copy-ratios ./out/DFG2.standardizedCR.tsv --denoised-copy-ratios ./out/DFG2.denoisedCR.tsv --sequence-dictionary WORKDIR/./RefGenome/dmel-all-chromosome-r6.39.dict --output ./out/plots --output-prefix DFG2

c) Entire program log:
I am using Drospophila samples to use gatk somatic CNV pipeline. The plot created by PlotDenoisedCopyRatios function turned out to be empty

When tracing back previous step files, I noticed that the DFG2.standardizedCR.tsv and DFG2.denoisedCR.tsv have entries like

CONTIG  START   END     LOG2_COPY_RATIO
2L      1       23513712        -15.784881
2R      1       25286936        -15.237691
3L      1       28110227        -15.292732
3R      1       32079331        -15.153327
4       1       1348131 -18.814539
X       1       23542271        -16.758953
Y       1       3667352 13.513223

where in the tutorial, https://gatk.broadinstitute.org/hc/en-us/articles/360035531092--How-to-part-I-Sensitively-detect-copy-ratio-alterations-and-allelic-segments, hcc1143_T_clean.denoisedCR.tsv file contain results such as

CONTIG  START   END     LOG2_COPY_RATIO
chr1    68839   70260   1.862435
chr1    925690  926265  -0.133840
chr1    929903  930588  0.114137
chr1    930787  931341  0.031698
chr1    935520  936148  0.127558
chr1    938788  939202  0.652190
chr1    939203  939712  -0.204635
chr1    940892  941558  -0.231103
chr1    942524  943155  0.321343

meaning the calculations should break the chromosome into parts rather than analyzing it as a whole.

Could you give me some idea of what could be wrong with the analysis? 
Prior to this step, I followed

PreprocessIntervals

CollectFragmentCounts

CreateReadCountPanelOfNormals

Then did the DenoiseReadCounts and PlotDenoisedCopyRatios

0

• 

  Gökalp Çelik

  • June 23, 2024 21:20

  Hi Yuwei Bao

  Can you provide us the first couple of intervals from the output of CollectFragmentCounts results?

  Can you also post your command line for each step so that we can check a bit more deeper?

  Regards. 

  0

  Comment actions Permalink
• 

  Yuwei Bao

  • June 25, 2024 14:39
  • Edited

  Hi Gökalp:

  Thank you very much for your response! I got it figured out. The step I did incorrectly was in the first step PreprocessIntervals. I want to share it here if others run into the same issue.

  $gatk_path PreprocessIntervals \
  
  -R $REF \
  
  --bin-length 0 \
  
  --interval-merging-rule OVERLAPPING_ONLY \
  
  -O $out/$SAMPLE.preprocessed.interval_list"

  
  Because I am working with non-human genome and don't have a interval list -L, if I keep the --bin-length 0, which will lead to the whole genome was not broken into parts. I was able to have PlotDenoisedCopyRatios work after customizing the --bin-length.
  
  During the exploration, I also run into an issue 
  

  java.lang.OutofMemoryError: GC overhead limit exceeded
  

  which can be fixed by increasing heap size such as 
  

  --java-options "-Xmx100g"

  
  Thanks for the great tool and help!
  
  Yuwei

  0

  Comment actions Permalink

Please sign in to leave a comment.