Can >3M deletions be detected by sniffles?

[yue131]

Hi,

Thank you for a great tool!
I have two patient samples, both of them have a deletion longer than 3M in chrY, and can be clearly visualized in IGV ((attached). To further compare the size of this deletion between two samples, I used Sniffes to call variantion, However this deletion was not included in the sniffes result.

below was my script:
sniffles -i mysample.sort.bam --vcf mysample.vcf --minsupport 2 --minsvlen 50 --long-ins-length 2000000 --long-del-length 5000000 --long-inv-length 5000000 --long-dup-length 5000000

And I had tried many ways to explain this result, such as add --no-qc or not, --minsupport 1, add --mosaic or not.

This is HiFi WGS data. Please help me to understand this.
Thank you
Best

[fritzsedlazeck]

Thanks @yue131 for reporting this. ChrY is always hard because of its unresolved nature (I assume thats Grch38). Its certainly not mosaic. The --no-qc should have called it if its visible. The issue might be that the alignments and coverage are not sending the same information.. We also have developed Spectre for large CNV calling.

Would it be possible to share this region with us over email?
Thanks
Fritz