Sib-pair Command: varcomp

ClassAnalysis and data manipulation command
Arguments <quantitative trait> [ae] [ce] [ace] [ade] [aqe (inline | ibs | <marker 1> + ...<marker N>)] [covariate <covariate trait 1> + ...<covariate trait N>].

Performs MVN variance components analysis for a quantitative trait or a binary trait using all phenotyped individuals. Currently fits ADE, ACE, AQE, AE and CE models. The ae option fits AE and E only, and so forth. Multiple covariates can be included in the fixed effects part of the model via adding the cov keyword and a list of covariates at the end of the command line. Only the first marker locus in the list of covariates is fully allelic effect coded, with subsequent markers included as the mean of their allele values (i.e. 1="1/1", 1.5="1/2", 2="2/2" for a diallelic marker).

The AQE variance components linkage analysis routines is usually accessed by the qtl...full command, but is also accessible by varcomp aqe. In the latter case, one specifies a set of markers to be combined to give a multipoint ibd estimate for the QTLs ibd status, or if the ibs modifier is present, then all active markers are used to estimate kinship from average marker ibs. One can also read a lower triangular matrix of ibd's for each pedigree from the command line, using aqe inline.

The multifactorial threshold model defaults to using the Mendell and Elston approximation to the MVN likelihood, but the Genz algortithm can be selected using the set mft command.


>> include
>> varcomp sace
# Examine effect of including ACE I/D genotype
# including allelic and then genotypic effects
>> varcomp sace covariates aceins
>> lrt
>> varcomp sace aqe aceins 
>> varcomp sace aqe aceins covariates aceins
>> set loc a12 qua
>> set loc a22 qua
>>  a12 = (aceins=="1/2")
>>  a22 = (aceins=="2/2")
>> varcomp sace covariates aceins
>> varcomp sace covariates a12 a22
>> lrt

# One hundred nuclear families with 
# three phenotyped offspring
simulate pedigrees 1000 2 3 3
set loc marker mar
set loc hitrait aff
set loc qtrait qua
sim marker 0.6 0.4
if (marker=="1/1") then qtrait=rnorm else qtrait=rnorm+1
let hitrait = (qtrait > 0)
if (isfou) then (hitrait=x) (qtrait=x)
des hitrait
set mft 10000 0.001
mft hitrait ae

Which gives:

Relative pair   RecRisk     Aff-Aff     Aff-UnA  Tetrachoric r
-----------------------------------------------  -------------
   Fullsib        0.725        1546        1172        0.087

NOTE:  Mendell-Elston MVN algorithm in use.

Multifactorial theshold model for "hitrait"
Random: A  

Number of families         =    1000
Number of MZ twinships     =       0
Number of observations     =    3000

Trait mean / intercept     =  -0.555285    
Trait prevalence           =   0.360534    

Additive genetic variance  =   0.171431     ( 17.1%)
95% Profile likelihood CI  =   0.858102E-01 --   0.264081    
Environmental variance     =   0.828569     ( 82.9%)
Model loglikelihood        =   -1801.11    

Chi-square testing VG=0    =       7.09     (df=1, P=0.008)
Total function evaluations =      87

See also:

mgt variance components allelic association
qtlpairsibs or pedigree VC linkage
lrt compare last 2 models fitted (mix/VC/GLM/GLMM)
fpmMCMC mixed/SML/finite polygenic
set mftMFT evaluation count and tolerance

<< (kendalltau)Up to index>> (lrt)