Manual for GDT

GDT is a software package including a variety of family-based genomewide association (GWA) tests. The main method GDT tests within-family association by examining differences of genotypes from all discordant relative pairs. The GDT method generalizes family-based association tests, in the sense that it extends methods such as TDT, 1-TDT, and PDT.


The input files include at least a data file, a pedigree file, and a map file. They can be specified with the -d, -p, and -m parameters. The command line will look like this:

prompt> gdt -d ex.dat -p ex.ped -m --gdt
GDT supports input files in LINKAGE, GENEHUNTER, and MERLIN format. Please read MERLIN tutorial for details of MERLIN input files.


Examples of association tests are:

prompt> gdt -d ex.dat -p ex.ped -m --stop 3 --cov sex --gdt
The output may look like this:
   Binary Trait Association : --TDT, --oneTDT, --PDT
   Quasi-likelihood Methods : --WQLS, --MQLS, --GEE1
           GDT & Extensions : --GDT [ON], --poGDT
   Quantitative Association : --fastAssoc, --robustAssoc
         Haplotype Analysis : --hapfile []
           Linkage Analysis : --ibdfile []
           Model Parameters : --trait [], --covariate [sex]
        Analysis Parameters : --prevalence
                  Positions : --start, --stop [3.00], --SNP []
                     Output : --prefix [GDT]

Fitted Polygenic Models (GEE)
          Trait     N       Mean        sex
            T1D 11674     -0.071     -0.151
           (se)            0.019      0.032

Phenotype: T1D [GDT] (2496 families)
Marker       Position   Pair Allele  Freq   Delta   Z_GDT  pvalue
rs741737     0.000000  10304      2 0.258   0.005   0.728    0.47
rs7481525    0.000050   9524      2 0.402   0.004   0.411    0.68
rs1105350    0.730000   4015      2 0.435   0.009   0.851     0.4
rs2029463    1.050000  10350      1 0.127   0.007   1.195    0.23
rs1004446    2.100000  10347      2 0.690   0.053   6.292 3.1E-10
In the above example, all markers with position < 3Mb are tested for association using the GDT method adjusting for SEX. There are 8 columns in the genome scan output: marker's name, marker's position, number of discordant relative pairs, the reference allele, frequency of the reference allele, allele frequency difference between cases and controls, the Z statistic, and P value.

All tests implemented in the current version include:
--TDT: the Transmission/Disequilibrium Test (Spielman et al. 1993)
--oneTDT: or 1-TDT, a TDT extension that allows one parent data (Sun et al. 1999)
--PDT: Pedigree Disequilibrium Test (Martin et al. 2001, 2002)
--MQLS: MQLS, more powerful Quasi-Likelihood Score method (Thornton and McPeek 2007)
--WQLS: WQLS (Bourgain et al. 2003)
--GEE1: a generalized estimating equations (GEE) implementataion assuming an independent working covariance (Liang and Zeger 1986)
--poGDT: a variation of GDT that only examines discordant parent-offspring pairs (Chen et al. 2009 AJHG 85:364-376)
--GDT: Generalized Disequlibrium Test (Chen et al. 2009 AJHG 85:364-376)


Methods implemented include:
--fastAssoc: the score test (Chen and Abecasis 2007)
--robustAssoc: robust score test (Chen et al. 2009 AJHG 85:364-376)


--trait trait_names specifies one or more traits to be analyzed. All traits will be analyzed if not specified.
--covariate covariate_names specifies covariates to be included in the model.E.g., --cov sex,age
--start position specifies a starting position for analysis
--stop position specifies a stopping position for analysis
--SNP SNP_names specifies a list of SNPs used in the analysis. E.g., --SNP SNP1,SNP2


Chen WM, Manichaikul A, Rich SS (2009) A generalized family-based association test for dichotomized traits. Am J Hum Genet 85:364-376 [PDF]

Chen WM, Abecasis GR (2007) Family-based association tests for genomewide association scans. Am J Hum Genetc81:913-926 [PDF]

Last updated: September 3, 2009 by Wei-Min Chen