I was just wondering for no particular reason. Am I genetically closer to my brother or to my son? Or is it the same?
In theory, if a parent gives no more than 50% of their chromosomes to a child, isn’t it possible (wildly unlikely, but possible) for siblings not to be all that closely related, DNA-wise? Mom could give half her genes to Kid One and the other half to Kid Two — both are her children, but both have different sets of stuff. Obviously, two brothers would share the same Y-chromosome, of course. Point being that you can share between some-really-low-number (0%?) & 100 (identical twin) % of your DNA with a brother, while your child is a guaranteed 50%, no matter how you slice it.
A related question I’ve always had: Humans have 46 chromosomes. We have 64 great-great-great-great grandparents. So logically we have at least 18 great-great-great-great grandparents that haven’t given us a single chromosome: are we related to them? mtDNA probably clouds this issue.
Chromosomes aren’t always passed on as a unit. They cross over, sections from paired chromosomes are exchanged.
In general chromosomes do not stay intact in transmission to offspring. The two paired chromosomes swap gene sequences so you likely have some genes from all your great to the n grandparents. And of course some genes are identical across many people so you couldn’t tell exactly which ancestor they’re from in any case.
The Y chromosome is an exception. It does not swap genes with the X so the father gives his Y chromosome to a son pretty much intact. So a son father and brother have the same Y chromosome pretty much.
The rest of the genes Apart from the X & Y are half from the mother and half from the father so a son has just about exactly half of his genes from his father. He will on average have about half the same genes as his brother but it could in theory vary from absolutely identical (even if not identical twins) to nothing but the Y in common.
Well, that’s pretty much been answered, so I’ll just do a summary:
So the theoretical overlap between brothers (excluding the Y chromosome) can be anything from 0 to 100 percent, though of course random distribution will hardly ever peg the meter one way or the other. But with the normal crossing over of material between chromosomes the practical range will be somewhat less than that.
The genetic overlap between father and son will always be 50 percent (less the difference in size between X and Y).
Slight nitpick, but the mother and father will of course share many genes in common, so in terms of genetic similarity (as in the OP), the genetic similarity will be much greater than 50% to either parent.
And further, if we are considering genetic similarity (and not where a particular gene was sourced from), it gets rather more complicated.
e.g. The parents have genotypes AA and BB for a particular gene.
All their offspring will have the genotype AB.
Which means that their children have a 100% chance of sharing genotype and phenotype with their siblings. But they have a 0% chance of sharing genotype with their parents.
Whether they’ll share either parent’s phenotype depends on which parent has the dominant allele. Furthermore if the alleles are co-dominant, then no child will share geno- or phenotype with either parent.
My intuition is that siblings are more similar, but I can’t crunch the numbers.
Nice question, diggleblop!
Actually, I’m wrong about this aren’t I? Not all chromosomes cross over. I think.
Think of it this way: one the one extreme are identical twin brothers, who share 100% of their DNA.
On the other extreme, think of two brothers. Brother A got all of his odd numbered chromosomes from his mother and all of his even numbered chromosomes from his father (or any other unit instead of chromosomes to avoid the crossing-over issue). The other brother just happened to get all of his even numbered chromosomes from his mother and all of his odd numbered chromosomes from his father, just through random chance. They share the same sex chromosomes - Y from their father and X from their mother - but otherwise they don’t share anything.
But this is quite an extreme scenario you’re proposing.
I’m not absolutely sure, and I’m waiting for a biologist to join this thread, but AFAIK, crossover is the rule, not the exception.
You inherit one allele from both parents for each gene. So, as I’ve already pointed out, if your parent’s genotypes are AA and BB, then the offspring will all have the same geno- and phenotype as their siblings but none will have the same genotype as their parents, and it’s quite possible that none will even have the same phenotype of either of the parents.
This confounds the “common sense” analysis that tells us that a given child is 0-100% the same as its siblings and 50% the same as its parents.
Of course it depends what the OP meant by “genetically closer”. I’d argue that defining it as a greater proportion of identical genotypes or phenotypes (how similar two people are genetically or biologically, respectively), makes a lot more sense than where the alleles were sourced from (which guarantees nothing and seems irrelevant).
I recall reading somewhere that one’s siblings are ones closest relatives, since, with the exception of each other, they have the exact same relatives. (i.e., same parents, brothers, sisters, cousins, aunts, uncles, grandparents, etc.)
Don’t know whether that helps the OP, but I thought I’d throw it in.
It’s also possible that they have different X chromosomes, as their mother has two.
So, on the average your brother shares about as many genes with you as your father and your mother. However, there’s much more of a possibility of variance with a sibling.
Does that sum it up pretty well? 
Yes, that sums up how genetics works in Imaginary Land very well.
I know I’m being crabby, but it seems no-one is listening to the point I’m trying to make.
:mad:
Let’s attempt to crunch through some of the numbers.
Imagine there’s a gene that has just two alleles: A and B. Forget dominance and all that, because we’ll look at the chance of the offspring sharing genotype with their parents only.
‘pParent’ = % chance same genotype as a random parent
‘pSibling’ = % chance same genotype as a random sibling
AA x AA = AA (pParent = 100, pSibling = 100)
AA x BB = AB (pParent = 0, pSibling = 100)
AB x AB = AA AB BB (pParent = 50, pSibling = 37.5)
AA x AB = AA AB (pParent = 50, pSibling = 50)
…and from this set of combinations it appears that a similarity to siblings is slightly more likely.
Of course it gets more complicated if we throw in extra alleles, codominance and so forth, but these extra cases almost certainly favour similarity to siblings anyway.
e.g.
AB x CD = AC AD BC BD (pParent = 0, pSibling = 25)
I just asked a very similar question to this one: DNA test for familial relationship. The answer I got there was that full siblings are just as closely related to each other as they are to their parents: 50%.
*** Ponder
What a shame…there was a geneticist on that thread, but he/she didn’t weigh in on the genetic similarity question.
In any case, even while we’re using the simplistic and wrong way of looking at genetics, it’s clear that you must mean 50% of the genes that vary among living humans.
If you’re just comparing the whole genetic code then I’d suppose the similarity of any two humans is above 99% (even if we include the Y chromosome).
Oh, and I realised that my table of genotypes didn’t include the fact that certain pairings e.g. AB x AB might happen more frequently than others e.g. AA AA.
So assuming the alleles A and B are equally distributed in the population.
From the set of permutations (not included) we see that:
The pairing AA x AA (or BB x BB) is 12.5% of couples
The pairing AA x AB (or BB x AB) is 50% of couples
The pairing AB x AB is 25% of couples
The pairing AA x BB is 12.5% of couples
So multiplying the probabilities of similarity to parent and sibling (from the table in my previous post), by the probability of that particular pairing, gives:
AA x AA = (pParent = 100 x .125 = 12.5) (pSibling = 100 x .125 = 12.5)
AA x AB = (pParent = 50 x .5 = 25) (pSibling = 50 x .5 = 25)
AB x AB = (pParent = 50 x .25 = 12.5) (pSibling = 37.5 x .25 = 9.375)
AA x BB = (pParent = 0 x .125 = 0) (pSibling = 100 x .125 = 12.5)
So for this particular gene, the chance of sharing the same genotype with a random parent = 12.5 + 25 + 12.5 + 0 = 50%
The chance of sharing the same genotype with a random sibling = 12.5 + 25 + 9.375 + 12.5 = 59.375%
Not a huge difference, but if all the genes favour similarity to sibling slightly*, then the chance of your entire genepool being more similar to a random parent than a random sibling is extremely, bet-your-life-on-it, low.
- And as I’ve already said, other situations such as more than 2 alleles, almost certainly favour sibling similarity, although there are too many permutations to list out.
As has been pointed out, the units being passed on are genes not whole chromosomes, so my question is, how many genes are there? And how many generations apart do you have to get before there is a greater than 50% chance that you have *no *genes inherited from your ancestor?
Thanks for the education in this thread. I can (sort of) narrow down my own question now.
The Human Genome Project suggests fewer than 30,000 protein-coding genes.
Going back fifteen generations, you have 32,768 great(x¹³)-grandparents. So, logically, they can’t have all given you genes.
Except that you’re bound to have a considerable number of repeat ancestors among those 32K.
And of course the point mentioned above about which of those genes are unique to humans.
Well, even though less than 5% of the genome codes for proteins, we’re now finding microRNAs have a huge contribution to protein expression. They call this branch the “Conductome”.
It’s kind of like the old belief that we only use 10% of our brain (which is incorrect). We use more than 5% of our genome, and it’s now thought that it might be a very high percentage. So you probably do have important regulatory components from all of those great grandparents, even if it’s not a protein.
Ok, my last word on the subject. I promise 
I think I can sum up my argument without recourse to a table of probabilities.
Like so:
Question: Who is genetically closer, father and son or brothers?
Answer: Almost-certainly brothers, though not by much.
Explanation: For every gene, each parent has two alleles, and each parent contributes one allele to form the pair that will determine the child’s genotype.
Because of this mechanism, the probability of brothers having the same genotype can never be less than 25%, regardless of the parents’ genotypes.
However, the probability of sharing the father’s genotype could be as low as 0%; if the father and mother have no alleles in common for that gene (e.g. Father has two A alleles, mother has two B alleles: all the children will have the genotype AB, which is different to either parent).
It’s this fundamental difference that skews the probabilities in favour of sibling similarity.
Greater similarity to the father is theoretically possible, but very unlikely, because if individual genes favour sibling similarity then it is overwhelmingly likely that the set of thousands of genes will also.