Autosomal recessive genes

[ann_h]

Can anyone tell me how you could work out the whole population incidence of carriers based on the annual birth incidence (#) of a disorder in its full form?

eg  for every child born with 2 defective genes, one will be clear and two will be carriers. There are, say, 800,000 people born each year ( x 70 years lifespan = 56 million population). If # is 1:10,000 then carriers are 2:10,000 or 160 per year. If # is 1:100,000 then carriers are 16 per year.

Or is there a geometric element of  probability to complicate it?

This isn't my homework, by the way!

[Peter Pedant]

Hi Ann

Your q is not very clear.

Incidence is, as you say, rate per year.

Prevalence is the flat how-many-there-are in a population at a certain time.

Arent you looking at the expansion of the Hardy Weinberg equation, p2 + 2pq + q2

if the prevalence of a disease is 1;400, (p2)

then the gene frequency is the square root of this, 1:20

Reading your question, I think this is what youa re getting at: I think youare asking if a disease has a frequency x what is the gene frequency, and the answer is the-square-root of x

For the reason given above.

[ann_h]

I think I mis-used terminology.

I know that the likelihood of a two-carrier couple producing an affected child is 1:4, with 2:4 carriers and 1:4 unaffected. This is standard. In this particular case, an affected child will not be able to have children, as they will die or be too disabled, so only the 2-in-4 that are carriers will count in my calculation of prevalence.

I have only been able to establish a rate of birth per year of affected children, so I am assuming that carriers will be double that, but maybe I am wrong.

Basically, if the birthrate is (x) per year, what are the odds of one carrier meeting another?

[Peter Pedant]

Hi Ann, there is nothing wrong with what you wondering, it is just that you havent got the words straight. You have used birthrate in a sense no-one else does.

Take a rarish disease like CF, prevalence 1: 400. Gene frequency 1 in 20.

Chance of two carriers meeeting 1/20 times 1/20.....and we are back to 1 in 400

Your point about lethality is not a factor so long as the disease is rare. If a gene hangs around for along time, even thugh the disease is lethal then it is called a founder effect - but that is not really what you're talking about, either.

If X knows he is a carrier, then the chance of meeting Y who is a carrier is 1/20 - conditional probability

Take something really rare like Ellis van Crefeld, 1 in 160 000, then the gene frequency is 1.400, and the upshot is that we never see it.....although oddly enough EVC disease shows a founder effect in the desert dwelling aborigines of Western Oz - which it shouldnt and the reasons for which are not really known

Does any of this help?

[ann_h]

Right then! Let's see if this is the right use of the words; then I can ask the question correctly!

Incidence = number/ratio of affected people born in any one year

Prevalence = number/ratio of affected people in the population as a whole

Gene frequency = number/ratio of people carrying the gene in the population as a whole

So, a gene frequency of 1:20 gives a prevalence of 1:400 by simple maths.

But, I only know the incidence (the annual number/ratio of births)

So what is the accepted principle by which gene frequency can be estimated from this?

(I didn't think it was the root as only a proportion of the population is of child-producing age, and they don't all have one baby every year)

Thanks for trying so hard, Peter, and sorry if I'm confusing you further!

Interesting about EVC, by the way.

[Peter Pedant]

The nice thing about AB is that one gets interested enough to ask someone who knows. One of my colleagues' wives is a geneticist, and I will ask.

The q is: If you know the incidence of a aut rec disease how can one estimate the gene frequency?

My suspicion is that you cant unless you know the number in the population and the average life span.

[Peter Pedant]

I tried this on a selection of my colleagues and they all came up with the first stab- square root of the incidence. They sort of agreed that you should nbeed the population, the average life span of the condition and the proportion of child bearing age -

gasp - the Lady who Knows, has been asked and we are breathlessly awaiting her answer!

[ann_h]

This is what I came up with, wait to see if they agree!

Incidence 1:90,000 of affected, who all die before procreating

Assume carriers therefore 2:90,000 or 22 per million

In a city of 1,000,000 people there are 22 carriers, of whom about 1 in 10 are in your age group.

So there's just the two of you.

[Peter Pedant]

The Lady who Knows hasnt answered. perhaps she thinks it is all too easy.

Incidence 1:90,000 - that is the number born per year remember. All dead by 15 say, so 16:90000 is the prevalence.

gene frequency 4:300

so in a city of however mmnay it is - quite a few more.

NOW - The nice thing about AB is finding outthe answer and so I wandered into a book shop. Try halliburton - theory of population genetic or something. p 77.

Estimation of gene frquency using the der- daaaah - hardy Weinberg equilibrium.

Anyway why are you doing this?

[ann_h]

It's such a rare disorder, I suspect somebody's skeleton just fell out of the cupboard

:(

[Peter Pedant]

A week later - the lady who knows er hasnt answered. I think I will have to give her a prod or something.

I havent got much further than reading Halliburton....

If you get to a sensible answer then please post it.

[ann_h]

I did find a website which seemed to agree with you, using the Hardy-Weinberg, except it came to 1:300 not 4 ( a square too far?!) It said there were some instances where you couldn't estimate. I'll have to find it again if you like.

As I said, it was a suspicion I had .