Re: Underestimating 'r'

Tim Tyler <tim@xxxxxxxxxxx> wrote or quoted:
> Perplexed in Peoria <jimmenegay@xxxxxxxxxxxxx> wrote or quoted:
> > "Tim Tyler" <tim@xxxxxxxxxxx> wrote in message news:dim7pm$q4$1@xxxxxxxxxxxxxxxxxxxxxx
> > > Perplexed in Peoria <jimmenegay@xxxxxxxxxxxxx> wrote or quoted:
> > > > "Tim Tyler" <tim@xxxxxxxxxxx> wrote in message news:dii3et$7nc$1@xxxxxxxxxxxxxxxxxxxxxx
> > > > > Perplexed in Peoria <jimmenegay@xxxxxxxxxxxxx> wrote or quoted:
> > > > > > "Tim Tyler" <tim@xxxxxxxxxxx> wrote in message news:dic6ai$sai$1@xxxxxxxxxxxxxxxxxxxxxx
> > > > > > > Tim Tyler <tim@xxxxxxxxxxx> wrote or quoted:

> > > > And why size of gene has any impact on the correct or measured value
> > > > of 'r'.
> > >
> > > The larger genes are considered to be, the smaller the probability of them
> > > being shared between individuals is - because of the greater probability
> > > of a difference occurring along their length.
> >
> > But the granularity of a gene makes practically no difference in the
> > value of 'r', whether you take 'r' to be defined by IBD or by regression.
> > Only McGinn would think otherwise.
> >
> > You seem to be suggesting that if we define a gene to be a big thing,
> > then we might estimate the average number of genes shared IIS between
> > random individuals to be 60% (say) but that if we define a gene to be
> > much smaller, then we would estimate the average number of genes IIS
> > between individuals to be much higher - 90% say. Yes, that is true,
> > but it doesn't change the value of 'r'.
> >
> > With big genes, the number of genes IIS between donor and a randomly
> > chosen member of the population may be 60%, but the number of genes
> > IIS between the donor and his full sib would be 80%. r = 0.5.
> >
> > With small genes, the number of genes IIS between donor and a randomly
> > chosen member of the population may be 90%, but the number of genes
> > IIS between the donor and his full sib would be 95%. It is still the
> > case that r = 0.5.
>
> Your figures are wrong :-|
>
> r - including r(regression) - changes as the gene size changes.

[snip one of my least-relevant examples in a long time :-|]

I'd like to take this opportunity to try to express my view again:

The idea that r(regression and IBD) is independent of the gene size in
an outbred population is largely true - provided the gene size is small
compared to the size of the bits of chromosome which are typically
preserved during crossover operations.

However, as this size of gene is approached, relatedness dwindles -
and as it is exceeded, relatedness drops down to near zero.

I think this independence also breaks down in inbred populations:

Inbreeding causes multiple crossover events to be relevant.

Chromosome stretches in the regions not shared by direct ancestry
are still identical by descent - but are only identical for
considerably shorter sections.

These regions effectively have an "r" of their own - and it is one more
likely to be influenced by gene size - since the identical regions are
likely to be shorter.
--
__________
|im |yler http://timtyler.org/ tim@xxxxxxxxxxx Remove lock to reply.

.

Relevant Pages

  • Re: Underestimating r
    ... >>> genes used in the derivation of Hamilton's rule. ... then we would estimate the average number of genes IIS ... the number of genes IIS between donor and a randomly ... Total genes: 20; Total identical: 9 ...
    (sci.bio.evolution)
  • Re: Underestimating r
    ... >> The larger genes are considered to be, the smaller the probability of them ... then we would estimate the average number of genes IIS ... the number of genes IIS between donor and a randomly ...
    (sci.bio.evolution)
  • Re: Underestimating r
    ... >> think that Williams's 'evolutionary genes' are bigger than molecular ... >> biology genes. ... And, more to the point, why they are bigger than the ... then we would estimate the average number of genes IIS ...
    (sci.bio.evolution)
  • Re: Underestimating r
    ... >>> With big genes, the number of genes IIS between donor and a randomly ... the number of genes IIS between donor and a randomly ... After reading what ...
    (sci.bio.evolution)
  • Re: Hamiltons rule
    ... The recipient receives a fraction r of his genes from the donor, ... >> and the remaining fraction from the general population. ...
    (sci.bio.evolution)