Re: Estimating temporal trends in additive genetic variance

Dear Matt

You wrote (complete text at bottom)
I wish to test whether the additive genetic variance of a trait is
changing over time. I have a study population where for the last 30 years,
all of the individuals born each year are measured for the trait at birth.
I have a pedigree for all individuals within the population.

I've set up a model as follows:

BYR is a 30 level factor for birth year included as a fixed effect to
control for temporal trends in environmental conditions.
sBYR is a standardized linear covariate of birth year centred on the mean
value and ranging from -1 to 1.
ANIMAL is the identity of the individual linked through the pedigree as a
pedigree factor (!P)
leg(sBYR,1).ANIMAL fits ANIMAL as a first order polynomial function of the
standardized birth year.

Trait ~ mu BYR !r leg(sBYR,1).ANIMAL

From this model I think I get estimates for the intercept, slope, and
covariance between the two from the random term:
int
cov slope

====
I'm sure someone else could offer a comment but here's may take.

The model term leg(),ANIMAL is typically used when there are repeated
observations on the animal,
but an animal is only born once so the data to estimate the terms is very
sparse.
You need two points at least to fit a slope but no ANIMALs have two
observations.

Furthermore, the A matrix accomodates a primary source of change in
genetic variation, that due to selection.
What other mechanism do you propose would change genetic variance over
time. Surelely it would
also change residual variance, so you would need to allow for that as
well.

You probably need a good number of years, which means a lot of animals
which means a big analysis,
Any change in variance that is systematic over time would need to be slow,
unless you have in mind
a particular change (say a change in technology measurung the trait at
some point in time)

If the heritability is low, it will be hard to detect any trend. If the
heritability is high, the trend may be evident
in the phenotypic variances calculated within years.

So, I would not be surprised if it proved very difficult, if not
impossible to fit such a model. But its probably worth a try.

May Jesus Christ be gracious to you,

Arthur Gilmour, His servant .

Mixed model regression mapping for QTL detection in experimental crosses.
Computational Statistics and Data Analysis 51:3749-3764 now available at
http://dx.doi.org/10.1016/j.csda.2006.12.031

Profile: http://www.dpi.nsw.gov.au/reader/17263
Personal website: http://www.cargovale.com.au/

mailto:Arthur.Gilmour_at_dpi.nsw.gov.au, arthur_at_cargovale.com.au
Principal Research Scientist (Biometrics)
NSW Department of Primary Industries
Orange Agricultural Institute, Forest Rd, ORANGE, 2800, AUSTRALIA

fax: 02 6391 3899; 02 6391 3922 Australia +61
telephone work: 02 6391 3815; home: 02 6364 3288; mobile: 0438 251 426

ASREML 2 is now available from http://www.VSNi.co.uk/products/asreml
The ASReml discussion group is at ASREML-L_at_dpi.nsw.gov.au
To join it, mailto:arthur.gilmour_at_dpi.nsw.gov.au
Cookbook: http://uncronopio.org/ASReml

Proposed travel:
Leave early September
AAABG 24-26 September Armidale
    <><><><><><><><><><><><><><><><><><><><><><><><><>
Dear all,

I was wondering if anyone had any thoughts on the following problem:

I wish to test whether the additive genetic variance of a trait is
changing over time. I have a study population where for the last 30 years,
all of the individuals born each year are measured for the trait at birth.
I have a pedigree for all individuals within the population.

I've set up a model as follows:

BYR is a 30 level factor for birth year included as a fixed effect to
control for temporal trends in environmental conditions.
sBYR is a standardized linear covariate of birth year centred on the mean
value and ranging from -1 to 1.
ANIMAL is the identity of the individual linked through the pedigree as a
pedigree factor (!P)
leg(sBYR,1).ANIMAL fits ANIMAL as a first order polynomial function of the
standardized birth year.

Trait ~ mu BYR !r leg(sBYR,1).ANIMAL

From this model I think I get estimates for the intercept, slope, and
covariance between the two from the random term:
int
cov slope

I can then run the model again changing the random term to
leg(sBYR,0).ANIMAL and compare the log-likelihoods to see whether the
additive genetic variance is changing over the study period.

I can also back transform the value I get onto the values of the
polynomial to get estimates of additive genetic variance for each birth
year and estimate the approximate standard errors using the matrix from
the vvp file, which I can then use to plot a graph of changing additive
genetic variance over time.

I see this as a potential way to test for the effects of selection acting
upon the trait to reduce additive genetic variance over time.

Would this work?

Thank-you all very much for your time.
Best wishes
Matt

*****************************************************
Matt Robinson

Institute of Evolutionary Biology
The University of Edinbrugh
Room 102, Ashworth Laboratories,
King's Building's Campus,
West Mains Road,
Edinburgh.
EH9 3JT

Tel: +44 (0)131 650 5462

e-mail: matthew.r.robinson_at_ed.ac.uk

This message is intended for the addressee named and may contain
confidential information. If you are not the intended recipient or
received it in error, please delete the message and notify sender. Views
expressed are those of the individual sender and are not necessarily the
views of their organisation.
Received on Sun Jun 20 2007 - 09:22:03 EST