No plots generated by the BaseRecalibrator walker

alexpenson

I cannot produce BQSR plots, although I can open the grp file with gsa.read.gatkreport.

Here's the command:

java -Xmx1g -jar $shares/GenomeAnalysisTK-2.3-6-gebbba25/GenomeAnalysisTK.jar \ -T BaseRecalibrator \ -I ./0.reorder.bam \ -R $shares/ftp.broadinstitute.org/bundle/2.3/hg19/ucsc.hg19.fasta \ -knownSites $shares/ftp.broadinstitute.org/bundle/2.3/hg19/dbsnp_137.hg19.vcf \ -BQSR ./0.reorder.bam.recal.grp \ -o ./0.reorder.bam.post_recal.grp \ --plot_pdf_file ./0.reorder.bam.post_recal.grp.pdf \ -L chr1:1-1000 \ -l DEBUG \ --intermediate_csv_file ./0.reorder.bam.post_recal.grp.csv

##### ERROR stack trace java.lang.NullPointerException at org.broadinstitute.sting.utils.Utils.join(Utils.java:286) at org.broadinstitute.sting.utils.recalibration.RecalUtils.writeCSV(RecalUtils.java:450) at org.broadinstitute.sting.utils.recalibration.RecalUtils.generateRecalibrationPlot(RecalUtils.java:394) at org.broadinstitute.sting.gatk.walkers.bqsr.BaseRecalibrator.generatePlots(BaseRecalibrator.java:474) at org.broadinstitute.sting.gatk.walkers.bqsr.BaseRecalibrator.onTraversalDone(BaseRecalibrator.java:464) at org.broadinstitute.sting.gatk.walkers.bqsr.BaseRecalibrator.onTraversalDone(BaseRecalibrator.java:112) at org.broadinstitute.sting.gatk.executive.Accumulator$StandardAccumulator.finishTraversal(Accumulator.java:129) at org.broadinstitute.sting.gatk.executive.LinearMicroScheduler.execute(LinearMicroScheduler.java:97) at org.broadinstitute.sting.gatk.GenomeAnalysisEngine.execute(GenomeAnalysisEngine.java:281) at org.broadinstitute.sting.gatk.CommandLineExecutable.execute(CommandLineExecutable.java:113) at org.broadinstitute.sting.commandline.CommandLineProgram.start(CommandLineProgram.java:237) at org.broadinstitute.sting.commandline.CommandLineProgram.start(CommandLineProgram.java:147) at org.broadinstitute.sting.gatk.CommandLineGATK.main(CommandLineGATK.java:91)

It looks like the csv file is not being produced.

Thanks!

Geraldine_VdAuwera

Alright, what's failing is the method that creates the data lines to be written to the csv file. It's an extremely simple operation so there's got to be something wrong with your data. Considering the interval that you're running on is extremely short, there may not even be any valid data at all in the part of the table that fails to write. You should try running again from the first step of recalibration with a much longer interval, eg 20:10000000-20000000.

Geraldine_VdAuwera

All the data needed for the plots should be in the gatkreport; what do you see in that file? Have you tried running without specifying the intermediate csv file?

alexpenson

The grp file looks good. I tried to use BQSR.R to make the plot, but it seems to require a csv file. If I specify the intermediate csv file, then it contains only the header line. If not, then it is not produced. Thanks

Geraldine_VdAuwera

What error do you get when you try using BQSR.R to make the plot?

alexpenson

After making some changes for ggplot2 v0.9.3: opts( -> theme( and theme_ -> element_

I get: Error in distributeGraphRows(list(a, b, c), c(1, 1, 1)) : object 'a' not found

It looks like a, b and c are created based on the lines in the csv file, and I don't have any.

Thanks Alex

Geraldine_VdAuwera

OK, that makes sense. Can you please run your original command again with -l DEBUG but without specifying the intermediate csv file filename, then post the console output? I need to know if with the internal default, the csv file gets created properly for you and if not, at what point it fails exactly.

alexpenson

Here is the output for the command above minus --intermediate_csv_file ./0.reorder.bam.post_recal.grp.csv Thanks!