Can Exomiser perform segregation analysis?

[nvadgama87]

Hi there,

I'm using Exomiser on a pedigree, which has a AD MOI.

However, is there a way for Exomiser to rank variants based not only on phenotype relevance, but whether the candidate variant is in all affected individuals?

Despite stating the phenotype in the .ped file, Exomiser sometimes considers variants in only the proband, which is absent in other affected individuals.

Is there a way for Exomiser to filter out non-shared variants in a pedigree?

Best wishes, Nirmal

[damiansm]

Hi Nirmal,

The default behaviour for Exomiser is to do segregation analysis exactly as you describe so there must be something wrong with your .ped or .yml configurations file I think. Do you have the .ped file location in the .yml and a correct multisample VCF. Do you have the affected status (=2) correctly stated in the .ped file. Maybe share your .ped file so we can check

Best wishes Damian

On Tue, Nov 8, 2022 at 8:07 AM nvadgama87 @.***> wrote:

Hi there,

I'm using Exomiser on a pedigree, which has a AD MOI.

However, is there a way for Exomiser to rank variants based not only on phenotype relevance, but whether the candidate variant is in all affected individuals?

Despite stating the phenotype in the .ped file, Exomiser sometimes considers variants in only the proband, which is absent in other affected individuals.

Is there a way for Exomiser to filter out non-shared variants in a pedigree?

Best wishes, Nirmal

— Reply to this email directly, view it on GitHub https://github.com/exomiser/Exomiser/issues/464, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHO4PBRRQUFCAZCK4NXWC3WHIC5VANCNFSM6AAAAAAR2AU2NA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[nvadgama87]

Hi Damian,

Thank you for your help. I've read several of your papers and am excited to implement Exomiser as a standard when analysing families. Just need to make sure I'm doing things correctly.

Attached is a fig showing the pedigree, and a .ped file I used. The first row is the proband (32_YJL_35). Interestingly, when I change the proband to another affected individual (in the yml file), the top variants are different, and mainly unique variants to that affected individual.

BW, Nirmal

[damiansm]

Hi Nirmal,

Let us look into it. The pedigree is definitely larger and more complicated than the ones we normally run Exomiser on but it should work. Have you tried setting the proband to 36_YJL_41? We would typically have the proband as one of the affected children. Exomiser relies on the Jannovar external library to perform the segregation so I am guessing a bit here and have reached out to my colleagues who developed Jannovar

Best wishes Damian

On Tue, Nov 8, 2022 at 1:17 PM nvadgama87 @.***> wrote:

Hi Damian,

Thank you for your help. I've read several of your papers and am excited to implement Exomiser as a standard when analysing families. Just need to make sure I'm doing things correctly.

Attached is a fig showing the pedigree, and a .ped file I used. The first row is the proband (32_YJL_35). Interestingly, when I change the proband to another affected individual (in the yml file), the top variants are different, and mainly unique variants to that affected individual.

BW, Nirmal [image: Family_1_ped] https://user-images.githubusercontent.com/51944346/200574484-0cff3909-3d71-4ba7-96e0-2beb4df6d8dd.png [image: Family_1_fig] https://user-images.githubusercontent.com/51944346/200574517-f3618001-82b8-4130-bd7b-dd9c917611ed.png

— Reply to this email directly, view it on GitHub https://github.com/exomiser/Exomiser/issues/464#issuecomment-1307206121, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHO4PAW7OAOIAXVFO3XSITWHJHEZANCNFSM6AAAAAAR2AU2NA . You are receiving this because you commented.Message ID: @.***>

[nvadgama87]

Hi Damian,

Thanks for chasing this up. I will try your suggestions.

I wonder if this may be due to the way I merged the VCFs. I used bcftools merge, which does not require a reference genome. Do you have any recommendations on how to merge the VCFs for Exomiser?

Best, Nirmal

[damiansm]

I think bcftools should be fine. I have only used vcftools but think the merge function should do the same thing. As long as the multisample VCF produced has all the samples described in the pedigree and each variant row has a genotype entry in the same order it will work

On Wed, Nov 9, 2022 at 7:29 AM nvadgama87 @.***> wrote:

Hi Damian,

Thanks for chasing this up. I will try your suggestions.

I wonder if this may be due to the way I merged the VCFs. I used bcftools merge, which does not require a reference genome. Do you have any recommendations on how to merge the VCFs for Exomiser?

Best, Nirmal

— Reply to this email directly, view it on GitHub https://github.com/exomiser/Exomiser/issues/464#issuecomment-1308329071, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHO4PDFXMEYOP73YWKMB5TWHNHGPANCNFSM6AAAAAAR2AU2NA . You are receiving this because you commented.Message ID: @.***>

[julesjacobsen]

I wonder if the issue here is possibly that Jannovar makes the assumption that the proband is always one of the leaf nodes in a pedigree? In other words, I'm not sure if it will consider children of a proband for the segregation analysis. This is only a suspicion so I'll need to verify this first. @nvadgama87 would that fit with what you see when you change the proband?