The TDT test
------------

^tdt^ varlist [weight] [,^p^ed^(^varname^)^ ^i^d^(^varname^)^ ^f^ather^(^varname^)^
        ^m^other^(^varname^)^ ^sex(^varname^)^ ^af^fect^(^varname^)^ 
        ^ac^ode^(^numlist^)^ ^e^min^(^varname^)^ ^pac^ode(numlist) ^mo^rigin ^po^rigin 
        ^ma^le ^fe^male ^rob^ust ^cl^uster(^n^uclear|^p^edigree|^t^riad) ^nowarn^]

The command requires two arguments - the variables containing the two alleles
at the locus considered. ^iweights^ are allowed

Description
-----------

This function performs a TDT for a specified (single) locus. The program 
determines which transmissions are informative and counts observed an expected 
transmissions of each allele in these. A chi-squared test on 1 df is 
calculated for each allele and a global chi-squared (on multiple df) is also 
calculated.  

In the presence of linkage, transmissions to more than one affected 
offspring in the same family may no longer be assumed to be independent. 
This can be dealt with by using the ^robust^ option. This option may also
be used to relax the assumption that the two parental transmissions to a 
single affected offspring are independent.  

Weighted analyses are used and, unusually, weights can be negative, thus 
allowing transmission of an allele less frequently than expected to be scored
positively in some circumstances. An important use of this concerns inheritance
of quantitative traits; we might define weights to be deviation of the 
offspring trait value from the population mean. Weights can also be used to 
exclude offspring that cannot be excluded in any other manner. Note that the 
weight expression uses ^offspring^ values of variables.

Note that alleles coded as zero are NOT treated as missing. Neither are zero 
pedigree member codes. If you want zero to be treated as a missing value, 
use ^mvdecode^ to recode them. This problem does not occur if you 
read in "preped" formatted files with the utility ^inprep^.


Options
-------

^ped^	  variable containing pedigree id (default is 1st variable)

^id^	  variable containing pedigree member id (default is 2nd variable)

^father^  variable containing id of father (default is 3rd variable)

^mother^  variable containing id of mother (default is 4th variable)

^sex^	  variable containing sex of subject (default is 5th variable)

^affect^  disease status variable indicating affected or not (default is 6th 
          variable)

^acode^   the code(s) indicating an affected individual (default is 2)

^emin^	  the minimum number of expected transmissions for which an allele will
	  be considered independently. Alleles rarer than this are aggregated 
	  with the next least common allele until this condition is met 
	  (default is 5).	

^morigin^ restrict analysis to transmissions from mother

^porigin^ restrict analysis to transmissions from father
	  (^morigin^ and ^porigin^ options cannot be invoked simultaneously)

^pacode^  restrict aanalysis to transmissions from parent with specified 
          values in the disease status field

^male^	  restrict analysis to transmissions to male offspring

^female^  restrict analysis to transmissions to female offspring
	  (^male^ and ^female^ options cannot be invoked simultaneously)

^robust^  protects against non-independence of transmissions within "clusters"
	  (see below)

^cluster^ specifies whether clusters should be regarded as ^p^edigrees,
	  ^n^uclear families, or ^t^riads (default is nuclear families)

^nowarn^  turns off diagnostic output concerning misinheritance or pooling of
	  low frequency alleles

Returned values
---------------

^tdt^ is an ^rclass^ command and returns the following values:

r(chi2)     (scalar) Global chi-squared test statistic
r(df)       (scalar) Degrees of freedom
r(p_val)    (scalar) p-value
r(chi2type) (macro)  "Score"

Examples
--------

. tdt allele1 allele2, emin(10) robust cluster(n)

. tdt allele1 allele2 [iw= diastolic - 80], acode(1 2)

See Also
--------

^gtdt^, ^mlgtdt^, ^inprep^

Author
------

David Clayton

Diabetes and Inflammation Laboratory		Tel: 44 (0)1223 762669 
Cambridge Institute for Medical Research	Fax: 44 (0)1223 762102
Wellcome Trust/MRC Building			david.clayton@cimr.cam.ac.uk
Addenbrooke's Hospital, Cambridge, CB2 2XY	www-gene.cimr.cam.ac.uk/clayton