NGS Data from multiple families

Question

[stevenlang123]

Hello ya'll,

I am investigating a rare single binary trait, I have collected NGS sequencing data from multiple families, some members of each family are affected, and some are not. It seems that although there are not many common variants, there are however areas of genes in which variation is common in affected individuals across families. How would you recommend going about analyzing this type of data?

Answer 1

[Kizuna]

It depends first of all on the mode of inheritance :

If your trait is inherited as AD:

I would advise you to remove all variants that have a MAF > 0.01% in the general population (e.g: 1KG and EVS), than go and prioritise the variants... give a high priority for stop, frameshift, splice site mutations. once you have a list of candidate variants, go for co-segregation in other family members. If you have a healthy brother/sister then your patient and his brother/sister should not share any candidate variants. If you have the parents and if your mutation is not de novo, then your patient and one of his parent should share the mutation.

If your trait is inherited as AR:

set the threshold MAF to >0.5% (1KG and EVS).. prioritise you variants. and then go for co-segregation: Lets say you found an interesting HMZ (M1:M1) variant in your patient: If you have the parents, then they should each carry one M1 allele (father: M1:WT mother : M1:WT).. If also you have a brother/sister unaffected, this sibling can be HTZ for M1 or WT for M1 but not M1:M1. If you have two affected members in the same family, so they both should be M1:M1.

-> the pathogenic variant should be : M1:M1 in your patient and his affected brother/sister and M1:WT in each of the parents..

Idem for compound HTZ (M1:M2). same MAF and co-segregation rules.

Hope this would help,

Kizuna

[stevenlang123]

Yes this does help thank you! Any programs you would recommend?

[Kizuna]

If you mean a program for co-segregation, I think there is no need:

Once you have the MAF of EVS and 1000Genomes in you patient vcf file, I would recommend to merge the .vcf files per family for example you can use cbind() in R than subset you data frame. You can also use excel.

So first filtering using the MAF threshold and than you go for co-segregation by choosing the genotypes in the family members..