Somatic SV calling without matched normal control

[Maximus Yeung]

Hello everyone,

Would there be a step-by-step instructions on how to call somatic SV using delly, if there is no match normal control, but there is a cohort of unmatch normal to create PON?

There is no concrete answer in this group, but just some vague qualitative description of potential methods.

Thank you.

[Tobias Rausch]

Hi,

If you have long-reads, I would recommend the CHM13 and pan-genome graph filtering that is now available in delly: https://github.com/dellytools/delly#alternate-alignments-for-genome-graphs

These alternate alignments are useful for rare diseases or tumor-only sequencing approaches.

Best, Tobias

--
You received this message because you are subscribed to the Google Groups "delly-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to delly-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/delly-users/56ce5a7b-3bc4-43b5-8e36-9a71accf078cn%40googlegroups.com.

[Tobias Rausch]

Hi,

For short-reads, a shortcut is to first call all normal samples:

delly call -g hg38.fa -o s1.bcf -x hg38.excl sample1.bam

then merge SV sites into a SV site list

delly merge -o sites.bcf s1.bcf s2.bcf ... sN.bcf

and use this site list as a panel-of-normal to filter "normal" SVs from a tumor sample with sansa: https://github.com/dellytools/sansa

sansa compvcf -y 0 -m 0 -n 250000000 -e 0 -a sites.bcf -o tumor tumor.bcf

In that comparison, the true positive (TP) sites are the "normal" SVs and the rest are candidate somatic SVs.

Best, Tobias













On Wed, Oct 23, 2024 at 11:00 AM Maximus Yeung <maxim...@gmail.com> wrote:

Hi Tobias,

Thanks for your reply.

For short reads NGS, may I know what the formal procedures using Delly to call tumour-only samples over PON?

I believe it will be:

delly call -g hg38.fa -o geno.bcf -x hg38.excl tumor1.bam control1.bam ... controlN.bam
delly filter -f somatic -o t1.somatic.bcf -s samples.tsv geno.bcf

If I have N controls, is that delly will call N times over the normal samples, every time I call SV on tumour-only samples?

Any ways that we can create a PON from N normal samples, and re-use later?

Thanks.

[Maximus Yeung]

Hi Tobias,

Thanks for your reply.

For short reads NGS, may I know what the formal procedures using Delly to call tumour-only samples over PON?

I believe it will be:

delly call -g hg38.fa -o geno.bcf -x hg38.excl tumor1.bam control1.bam ... controlN.bam
delly filter -f somatic -o t1.somatic.bcf -s samples.tsv geno.bcf

If I have N controls, is that delly will call N times over the normal samples, every time I call SV on tumour-only samples?

Any ways that we can create a PON from N normal samples, and re-use later?

Thanks.

On Wed, 23 Oct 2024 at 15:20, Tobias Rausch <rausc...@gmail.com> wrote: