MI-TDT - Multiple Imputation for Transmission Disequilibrium Test.


MI-TDT [options] file.ped


MI-TDT is an application based on a Multiple Imputation algorithm which allows dealing with missing information in trio data (one affected child and both parents). Missing information could be understood as missing haplotypic phase or missing genotype.


Option for both algorithms:

-H <n> <n> is 0 or 1, for reconstruction under H0 or H1. Default: 0.

Options for EM algorithm:

--step <n> number of SNPs added at each step. Default: 2. If the program is too slow, you may try --step 1
--th <t> threshold under which familial configurations are pruned. Default: 1e-2.
--maxconf <n> maximum number of familial configuration. Default: 1000.
--EMth <t> threshold under which an haplotype frequency is set to 0 during the EM run. Default: 1e-10.
--EMcv <t> Convergence criterion for the EM. Default: 1e-3.

Options for IP algorithm:

--nfile <n> Number of imputed files. Default: 10.
--niter <n> Number of IP iterations before each imputed file. Default: 200.
--alpha <x> The initial value of the data augmentation parameter. Default: 0.2*(number of families in the input file)
--base <name> Basename for imputed files (name.1, name.2, etc) Default: the name of the input file.


This manual was written for MI-TDT version 0.1.