ITDseek

Detect FLT3 internal tandem duplication (FLT3 ITD) from amplicon sequencing reads (ITDseek) and simulate these reads (ITDsim)

Download latest version in a ZIP package

FLT3 ITD detection performance of ITDseek, Pindel, GATK HaplotypeCaller and Samtools based on 40,401 simulated samples of ITD length up to 201bp at different positions of an amplicon (Au et al. Diagnostic Pathology 2016 11:11 Figure 2)

ITDseek for detection

Requirements, as tested on 64-bit CentOS 5.5

• SAMtools version at least 1.3 (to support depth >8000x)

Usage

./itdseek.sh <sample.bam> <ref.fasta> <samtools> > sample.itdseek.vcf

• sample.bam: indexed BAM alignment file, generated from BWA-MEM with -M option (Mark shorter split hits as secondary). To index the BAM file: samtools index sample.bam
• ref.fasta: indexed hg19 reference genome in FASTA format. Only positions chr13:28607161-28609590 will be considered, or itdseek.sh needs to be modified accordingly. To index the FASTA file: samtools faidx ref.fasta
• samtools: path to samtools executable. Use samtools if it is already included in a directory defined in $PATH
• standard output (STDOUT): variant calls in VCF version 4.1

Examples

# Provided example data
# 100bp ITD leading to soft-clipping in BAM file
./itdseek.sh examples/ITD.100.100.bam ucsc.hg19.fasta samtools > ITD.100.100.itdseek.vcf
# 20bp ITD leading to insertion in BAM file
./itdseek.sh examples/ITD.100.20.bam ucsc.hg19.fasta samtools > ITD.100.20.itdseek.vcf

# Your own data
# BWA-MEM with -M
bwa mem -R '@RG\tID:ITDsample\tSM:ITDsample' -M ucsc.hg19.fasta ITDsample.R1.fastq ITDsample.R2.fastq | samtools sort > ITDsample.bam
# Index BAM
samtools index ITDsample.bam
# ITDseek
./itdseek.sh ITDsample.bam ucsc.hg19.fasta samtools > ITDsample.itdseek.vcf

Interpretation of results

Quality score of called variants is defined as the total number of sequencing reads with ITD detected (forward and reverse combined). The following is ITDseek output for a simulated 100bp ITD dataset included in the folder examples/, with overall depth 2000X and VAF 50%. The quality score (sixth column QUAL) is 1000, corresponding to variant depth 1000X. Additional details are described in the last column (eighth column INFO), including forward and reverse ITD reads DP2, ITD length LEN and ITD sequence SEQ.

$ ./itdseek.sh examples/ITD.100.100.bam ucsc.hg19.fasta samtools
##fileformat=VCFv4.1
##source=ITDseekV1.2
##reference=file://ucsc.hg19.fasta
##INFO=<ID=DP2,Number=2,Type=Integer,Description="# alt-foward and alt-reverse reads">
##INFO=<ID=LEN,Number=1,Type=Integer,Description="length of ITD">
##INFO=<ID=SEQ,Number=1,Type=String,Description="sequence of ITD">
##INFO=<ID=CLIPPING,Number=0,Type=Flag,Description="ITD is detected as soft-clipping">
##INFO=<ID=INSERTION,Number=0,Type=Flag,Description="ITD is detected as insertion">
##INFO=<ID=VAF,Number=1,Type=Float,Description="ITD allele fraction">
#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
chr13	28608310	.	G	GATTCTTACCAAACTCTAAATTTTCTCTTGGAAACTCCCATTTGAGATCATATTCATATTCTCTGAAATCAACGTAGAAGTACTCATTATCTGAGGAGCCG	2000	.	DP2=1000,1000;LEN=100;SEQ=ATTCTTACCAAACTCTAAATTTTCTCTTGGAAACTCCCATTTGAGATCATATTCATATTCTCTGAAATCAACGTAGAAGTACTCATTATCTGAGGAGCCG;CLIPPING;VAF=0.50

A simple filter based on quality score (i.e. number reads supporting ITD) is recommended for real experimental datasets, for example:

# Minimum quality score 50
./itdseek.sh ITDsample.bam ucsc.hg19.fasta samtools | awk '$1 ~ /^#/ || $6 >= 50' > ITDsample.itdseek.vcf

ITDsim for simulation

Requirements, as tested on 64-bit CentOS 5.5

• BioPerl (version 1.6.901 tested)

Usage

./itdsim.pl | ./itdsim2fastq.pl

This generates the evaluation dataset described in the manuscript in the current directory, specifically 40401 pairs of FASTQ files (2x275bp at 2000X depth and ITD VAF 50%) representing 201 startings positions and 201 ITD lengths for the amplicon "FLT3.ITD.line.29.chr13.28607916.28608351_tile_2" in Illumina TruSight Myeloid Panel. Please refer to the comments in the source code of itdsim.pl and itdsim2fastq.pl for viewing/modifying simulation parameters.

Citation

Au CH, Wa A, Ho DN, Chan TL and Ma ESK, 2016. Clinical evaluation of panel testing by next-generation sequencing (NGS) for gene mutations in myeloid neoplasms. Diagnostic Pathology 11:11 (doi:10.1186/s13000-016-0456-8)