ATAC seq - Odd insert size distribution

Question

[emalekos]

Hello, I'm relatively new to working with ATACSeq data. One of the major QC checks is to plot the insert sizes after mapping. In one of my ten samples I am seeing an unusual distribution, while the rest appear to look reasonable (single representative of the other nine below).

Any ideas on what might cause this/any suggested actions?

mapping command:

bowtie2 --very-sensitive --maxins 1000 -x ${bowtie2Index} -1 ./${dir}/${prefix}_1.fastq.gz -2 ./${dir}/${prefix}_2.fastq.gz

histogram command:

java -Xmx20g -jar picard.jar CollectInsertSizeMetrics I=${name}.bam H=${name}_histogram.pdf

All samples mapped at >95%