dindel stage 4 for pooled samples

[bgu]

How to set the --numBamFiles and --numSamples parameters for the fourth step in a dindel analysis of pooled samples.

I have 27 individuals sequenced. Data for each individual is contained in one BAM file. I have successfully completed the first three stages for the 27 BAM files. I now have 807 glf files for each of the 27 individuals.

However, how call the mergeOutputPooled.py script? Specifically what are the appropriate values of --numSamples and --numBamFiles. If I set the latter to the number of BAM files actually analyzed (27) the program quickly crashes. Running the program with --numSamples 27 and --numBamFiles 1 seems to work but is this the right thing to do? The parameters appear to be largely undocumented. In the end the run with --numBamFiles 1 ended with a vcf file that only contained a header and no results.

The actual error message was:

Reading /.../JERCANM000000.../dindel/stage3.100.glf.txt
An error occurred!
Traceback (most recent call last):
File "/opt/.../mergeOutputPooled.py", line 620, in <module>
main(sys.argv[1:])
File "/opt/.../mergeOutputPooled.py", line 613, in main
processPooledGLFFiles(glfFilesFile = options.inputFiles, maxHPLen = options.maxHPLen, refFile = options.refFile, outputVCFFile = options.o utputFile, doNotFilterOnFR = (not options.filterFR), filterQual = int(options.filterQual), numSamples = int(options.numSamples), numBamFiles = int(options.numBAMFiles))
File "/opt/.../mergeOutputPooled.py", line 336, in processPooledGLFFiles
raise NameError('Inconsistent glf files! Is the number of BAM files correctly specified?')
NameError: Inconsistent glf files! Is the number of BAM files correctly specified?

[meltee]

I've been attempting to run a similar analysis and have encountered problems as well. Have you had any success yet?

There is an option for stage 3 where you can substitute the --bamFiles option for the --bamFile. Instead of running each bam file individually, you provide a list of your bam files (with paths). You will then be able to run stage 4 with the correct number of bam files specified.

I've been able to complete the analysis to this point and call candidate indels. However, the VCF file produced does not contain genotypes for my samples. I'd rather not have to go back and call genotypes individually. Does anyone have any insight on how to get dindel to give the genotypes for all samples with the -doPooled option?

Any insight anyone has would be greatly appreciated.

[bgu]

No reply - bad sign?

Hi,

There has been no response. I suspect that running on each BAM file separately was a mistake. I guess I should have run for each interval on all BAM files simultaneously. It seems that all the computer time (~6½ cpu years) was spent for nothing

However this is not clear from the documentation.