Protocol for finding variants in GATK

[student-t]

Hi,

I'm looking for a way to use GATK for finding variants. Unfortunately, their documentation is poor and misleading. The commands that they present on their presentation slides are out-of-dated.

Can anyone post a simple working script or protocol for GATK given a raw BAM file?

[WhatsOEver]

What documentations are you referring to?
Although I agree that information is sometimes hard to find on their website, I personally think you'll have a hard time to find something with a better documented workflow than "GATK variant calling" in the world of bioinformatics.
Have a look at this site which contains highly detailed presentation slides on how to use GATK from Sep 2014. https://www.broadinstitute.org/gatk/...ations?id=4768

[student-t]

I have read the documentation but still having problems. What I'm doing now is that, I'm simulating variants, mutations etc from a reference genome. I'd expect I'd see something from the GATK tool. I ran through the pipeline but my resulting VCF file contains headers but no data. The problem is that I don't know whether I've screwed up my data or the way I'm using GATK is simply wrong. If there was a known protocol, then I could compare my steps with it.... Can anybody take a quick look at what I've done? It's paired-reads.

Here's my pipeline:

- silico.fa is my reference genome



- Create Index for FASTQ

samtools faidx silico.fa

- Generate a sequence dictionary

java -jar ../../Tools/Picard/picard.jar CreateSequenceDictionary REFERENCE=silico.fa OUTPUT=silico.dict

- Alignment

bwa mem -t 20 -M -R '@RG\tID:group1\tSM:sample1\tPL:illumina\tLB:lib1\tPU:unit1' silico_index reads/simulated_1.fastq reads/simulated_2.fastq > aligned.sam

- Sort SAM to BAM

java -jar ../../Tools/Picard/picard.jar SortSam INPUT=aligned.sam OUTPUT=sorted.bam SORT_ORDER=coordinate

- Mark Duplicates

java -jar ../../Tools/Picard/picard.jar MarkDuplicates INPUT=sorted.bam OUTPUT=marked.bam METRICS_FILE=metrics.txt

- Build BAM Index

java -jar ../../Tools/Picard/picard.jar BuildBamIndex INPUT=marked.bam

- RealignerTargetCreator

java -jar ../../Tools/GATK/GenomeAnalysisTK.jar -T RealignerTargetCreator -R silico.fa -I marked.bam -o realigner.intervals

- IndelRealigner

java -jar ../../Tools/GATK/GenomeAnalysisTK.jar -T IndelRealigner -R silico.fa -I marked.bam -targetIntervals realigner.intervals -o realigned.bam

- Haplotype Analysis

java -jar ../../Tools/GATK/GenomeAnalysisTK.jar -T HaplotypeCaller -R silico.fa -I realigned.bam --genotyping_mode DISCOVERY --heterozygosity 0.01 --defaultBaseQualities 30 -o haplotype.vcf

[dpryan]

The question becomes how the simulated reads were made.