GATK indel settings help please

[inou13]

Dear all,
I am new to GATK and try to test this to look at indels from ILLUMINA nextGen data generated from custom pull down experiments. I have 2 set of files, one "disease" file and one "no-disease"( from the same individual). I know from other experiments that there is a 4pb duplication in the disease sample, I found it as well using SAMTOOLS pileup .

I wanted to try GATK on the same set of samples and get indels using SomaticIndelDetector.
I have tried without local realignment first
java -Xmx2g -jar GenomeAnalysisTK.jar \
-R ref.fasta \
-T SomaticIndelDetector \
-o indels.vcf \
-verbose indels.txt
-I:normal normal.bam \
-I:tumor tumor.bam

from the output vcf I cannot detect my duplication, I have tried as well in the single file mode (without putting a normal reference file).

I tried to realign using both RealignerTargetCreator and IndelRealigner. It appears that this region is called in the RealignerTargetCreator as it is a bit complex.
I used this for my disease file and did the same thing for the normal file
java -Xmx2g -jar GenomeAnalysisTK.jar \
-I input.bam \
-R ref.fasta \
-T RealignerTargetCreator \
-o forIndelRealigner.intervals \

then I ran IndelRealigner for both files, using the just created forIndelRealigner.intervals .
java -Xmx4g -jar GenomeAnalysisTK.jar \
-I input.bam \
-R ref.fasta \
-T IndelRealigner \
-targetIntervals intervalListFromRTC.intervals \
-o realignedBam.bam \

With my new realigned Bams I did re-run SomaticIndelDetector , and did not detect any duplication, this region is not called. I have tried as well in the single file mode (without putting a normal reference file), no calls.

I have checked that this region is not discarded ( as containing too few or too much reads ) ..

Could you please help/advise for better settings?
Many thanks

[RDW]

With my new realigned Bams I did re-run SomaticIndelDetector , and did not detect any duplication, this region is not called. I have tried as well in the single file mode (without putting a normal reference file), no calls.

Have you also tried the UnifiedGenotyper in indel mode on the realigned 'disease' file? (I think the SID is basically the old Indel Genotyper V2, and I get the impression they only keep it in GATK because it has the somatic mode that the UG lacks).

[inou13]

HI, thanks for your tip, I can find an indel now 5 pb near where my indel is supposed to be, so somewhat not the sequence and position I expected ... could it be due to the realignment?
this is the command line I have used for UG
-jar GenomeAnalysisTK.jar -R /l.fa -T UnifiedGenotyper -l INFO -glm INDEL -o indels_UG.vcf -I /realigned.bam
thanks