Re: Mutation classification matrix for dnenrich

[Menachem Fromer]

Hi Chloe,

A good question.

We defined LoF (loss of function) as including nonsense (stop gained),
frameshift indels, and essential splice site mutations.

We defined NS (non synonymous) as including all non-silent mutations,
including missense, all LoF types, but all other coding changes as
well, including stop-lost (readthrough) and in-frame
insertions/deletions (any indel with CDS overlap length that is a
multiple of 3).

The "silent" mutation class consists of the set of synonymous mutations.

Best,
Menachem

> On Sep 22, 2015, at 2:16 PM, Chloe O'Connell <chloe.p....@gmail.com> wrote:
>
> Dear Dr. Fromer,
>
> I hope this email finds you well. I had a quick question about the mutation classification matrix for dnenrich, as I'm trying to convert VEP annotations into the categories you list.
>
> What is the distinction between loss of function and nonsense mutations, as well as nonsynonymous and missense mutations? I am having a bit of trouble figuring out how to classify a mutation in which a stop codon is lost (among other mutation types, such as synonymous mutations, inframe indels, and stop codon gains).
>
> I've reviewed the documentation on the site, but haven't been able to figure out how to map my existing annotations onto your variant classifications, so any help you can provide would be much appreciated.
>
> Thanks so much,
>
> Chloe O'Connell

[Chloe O'Connell]

Dear Dr. Fromer, 

Thank you for your quick reply. I guess I am still a bit confused about your classification of NS mutations, given that there is also a separate classification for "missense" mutations, but in your explanation it seems contained in the definition of non-synonymous. Is the classification of "missense" actually used, and if so, what is the distinction between it and NS? 

I've attached a shareable link to the conversion I've been working on just in case anyone else is would like to use VEP annotation in dnenrich in the future. It is a work in progress, and I will update as I get more information: https://docs.google.com/spreadsheets/d/1nYZSqlk0afK_Khj__0s4Wk1hIes5QJBa9kkJWS6dbNI/edit?usp=sharing (feel free to comment with suggestions/corrections)

[Menachem Fromer]

Indeed, missense IS INCLUDED in the NS category.

What you exactly decide to test is somewhat of a subjective judgement call and possibly dependent on the disease being studied and its negative selection, for example.

In our schizophrenia study, we focused on the narrow class of most deleterious LoF mutations and then expanded outward to the broader more inclusive class of NS mutations (including all coding, non-silent mutations).

We mainly did this to increase power (by combining the missense and LoF in a single class), but you could just as easily envision running LoF and missense as the 2 main categories.

Another advantage of NS is that includes other coding variants that are not missense or LoF, including stop lost (readthrough) and in-frame indels (which are often equivalent to one or 2 missense mutations).

Best,

Menachem