Genetic Distance (for discussion)

ChrisBowyer · Post by **ChrisBowyer** » 29 Sep 2006 09:24

Is this a new idea or has it already been described? Have you ever received an e-mail from someone and thought 'Who on earth are these people and why do I care what their grandparents were called? I think the answer is, it depends how closely related you are, and I was looking for a way of measuring that (rather than all those clumsy phrases like second-cousin-twice-removed or great-aunt's mother-in-law) which I find difficult to compare.

My proposal is what I'd like to call a Genetic Distance... Starting with me at as factor 1, I share half of each of my parent's genes, so they are factor 2 (i.e. statistically one in two genes in common). My brother is also a 2 since he has half my mother's and half my father's genes, so for any specific gene there's a one in two chance we have the same one. Grandparents are obviously 4, as I think are first cousins (though my genetics, as opposed to my gegealogy is gets stretched about this point), and it should be possible to find a way of calculating a similar factor to any relation on the tree.

The numbers you get go 1, 2, 4, 8, 16, etc which is not very linear, but if I want a 'distance' rather than this factor to compare them, someone with more maths than me can work out the function to convert it into 0, 1, 2, 3, 4, etc. and of course it could have fractions for half-siblings etc. (It wants to start at zero because it's a 'distance' from me.)

If anyone apart from me thinks this is not completely barmy, maybe we could lobby to get the genetic distance between people as some kind of function in FH. Perhaps use it in queries, or to set flags for icons, or whatever. And if we ever get the ability to identify a root individual in the records window, we could even sort people according to how closely related they are.

Any thoughts?

ID:1870

YouTube · Post by **Jane** » 29 Sep 2006 09:33

Is this not similar to the Code version of the =relationship function?

ChrisBowyer · Post by **ChrisBowyer** » 29 Sep 2006 10:02

Errr... no idea... what's that?

(Sorry)

YouTube · Post by **Jane** » 29 Sep 2006 10:38

Function Name: Relationship

Description: Provides information about how person B is related to person A.
CODE A coded description which gives more detail about how the two people are related. This description will have a format, something like the following: 'Gens. Up=2, Gens. Down=1, Half=0, Spouse Start=0, Spouse End=0'.
These are the same values that can be returned individually, by using the 'RelationCode' function. See that function for more detail about the meaning of these numbers.

The above is lifted from the Functions section of the online help, combined with the Relation Code I think you can see pretty well what you are looking for.

ChrisBowyer · Post by **ChrisBowyer** » 29 Sep 2006 10:48

Thanks Jane, I'll look into that

ejp · Post by **ejp** » 29 Sep 2006 11:39

I was reading about this the other day, sparked by the wish list item 98 - compute genetic relationship. This is one of the articles I read which refers to 3 ways of describing it.
http://www.bellanti.org/cousin_and_othe ... hips__.htm
I often wonder about this myself. Who is MORE related once out in the branches?

ChrisBowyer · Post by **ChrisBowyer** » 29 Sep 2006 12:13

Interesting article, thanks for pointing me to it... It comes to almost exactly the same figure for 'degree of kinship' as I did (except I forgot about the descendents of married cousins, not to mention identical twins, and the like which make it horribly complicated)

Tombaston · Post by **Tombaston** » 30 Sep 2006 09:43

There is already a wish list item 'Compute Genetic Relationship' for this as well as 'Graphical Relationship Calculator' which I feel could be more useful.

One problem with calculating the genetic relationship is it assumes an exact 50% split of genes. This suggests I have exactly 50% of the same genes as each of my two brothers. In reality I may have 44% the same as one and 53% the same as the other. Also It assumes I have 25% of the genes of each of my grandparents. Again it may be that I have only 23% from my paternal grandfather and 27% from my paternal grandmother. Also I know that I have at least one gene in common with all my direct male ancestors as well as all my male relations of whatever distance who are also direct male descendants of anyone on my direct male line. The only genetic relationship I can be sure of is that I have exactly 50% of my father's genes and 50% of my mother's. As the relationship gets further than this then errors in the split are introduced which will get bigger the further away the relationship is.

ChrisBowyer · Post by **ChrisBowyer** » 30 Sep 2006 10:26

Fair point Dave... however it's done it's only going to be a statistical probability, but still of some value I think. I hadn't spotted the wish list item... I think the degree of kinship figure is probably more useful than a percentage, but it that's not too important. The significant thing is it's a number that you can compare and sort by, and just get a feel for where you are when you're in the outlying reaches of the tree. I don't envy whoever has to work out how to calculate it for the descendents of related partners though.

Tombaston · Post by **Tombaston** » 30 Sep 2006 14:54

Understood.

I am in occasional correspondence with one of my wifes sixth cousins twice removed, and it has taken me five minutes to check that degree of relationship. Although my wife doesnt feel any great link to her distant cousin, they do have a common ancestry and it is nice to know how close or distant they are related when we correspond.

Family Historian User Group