==========================================
PRACTICAL EXERCISE FOR ASSOCIATION SESSION
==========================================



In this exercise, we will examine the pattern of linkage and
association between the angiotensin-1 converting enzyme gene on 
chromosome 17 and the circulating angiotensin-1 levels in blood.

This data-set was originally described by:

Keavney B, McKenzie CA, Connell JM, Julier C, Ratcliffe PJ, Sobel E, 
Lathrop M, Farrall M. (1998) Measured haplotype analysis of the
angiotensin-I converting enzyme gene. Hum Mol Genet 7:1745-51.

We will be working on standardized trait levels.

COPY THE FILES TO YOUR LOCAL COMPUTER
=====================================

We will use three files in this exercise:

keavney.dat 
keavney.map
keavney.ped 

You can find these at h:\goncaloa\angio

EXAMINE THE FILES
=================

These are all text files you can examine their contents directly or
you can run pedstats for a summary of the genetic information they
contain.

> pedstats -d keavney.dat -p keavney.ped

CALCULATE THE IBD DISTRIBUTION
==============================

We will use MERLIN for this step. To run merlin type:

> merlin -d keavney.dat -p keavney.ped -m keavney.map --ibd --marker

Other interesting options for this data set would be:

--dev      Calculate a simple non-parametric linkage statistic
           assuming phenotypes are deviates with mean zero.

--best     Haplotype these families, place the results in merlin.chr.

ESTIMATE LINKAGE
================

We compare a model including only residual environment and polygenic 
variances with one including a linked major locus. In QTDT, these 
variance components are labelled e (non-shared environment), 
g (polygenes) and a (major gene).

We will not model association at all (a-) and will use the IBD information estimated by MERLIN.

To specify alternative models for the variances use the -w and -v options in QTDT.

>qtdt -d keavney.dat -p keavney.ped -i merlin.ibd -a- -weg -vega

Examine the output. In particular, you should find:

* A description of the models being tested.
* Likelihood ratio chi-squared statistics at each marker.

Q: Do you think you could localize this gene by linkage?


ESTIMATE WITHIN FAMILY ASSOCIATION
==================================

Assuming an environmental, polygenic and major gene component for 
the variances, we will test for association using the between-within
model.

>qtdt -d keavney.dat -p keavney.ped -i merlin.ibd -ao -wega

Notice how all likelihood ratio chi-squared statistics for
association are much greater than in the linkage test.

Q: Do you think that association is helping localize the gene?



ESTIMATE LINKAGE AFTER MODELLING ASSOCIATION
============================================

We will use a similar test to the linkage test, but also model 
association on the means.

>qtdt -d keavney.dat -p keavney.ped -i keavney.ibd -ao -weg -vega

Q: How is this sort of information helping localization?
Q: Do you feel it would help design further studies?