QTDT - Twins and Other Multiples

Special care is required when analysing twin data. Twin status is specified using a zygosity column in the pedigree. QTDT assumes that all MZ twins have the same genotype and sex and will accurately estimate kinship and IBD coefficients for monozygotic twins. For IBD calculations carried out using an external program monozygotic twins are merged into a single individual.

Twin status is specified through a zygosity column in the pedigree file. The presence of this column must be indicated in the data file. Monozygotic twins are indicated using the MZ code. Dizygotic twins used the DZ code. For sibships will more than one set of twins a different odd (even) number should be used for each set of monozygotic (dizygotic) twins. The zygosity column should include a zero value for all individuals who are not twins.

For example, the following data and pedigree files specify that:

1. In family 1, individuals 3 and 4 are MZ twins
2. In family 2, individuals 3 and 4 are DZ twins
3. In family 3, there are two pairs of twins: individuals 3 and 4 are MZ twins, individuals 5 and 6 are also MZ twins. Since there are multiple twin pairs in this sibship numeric codes have been used, in this case odd numbers indicate that the twins are MZ.

 Z twin_status
 M growth_factor_snp
 T height

 1  1  x  x  1   0   1/2  160
 1  2  x  x  2   0   1/2  150
 1  3  1  2  1  MZ   2/2  100
 1  4  1  2  1  MZ   2/2  102
 2  1  x  x  1   0   1/2  160
 2  2  x  x  2   0   1/2  150
 2  3  1  2  1  DZ   1/1   90
 2  4  1  2  1  DZ   2/2  110
 3  1  x  x  1   0   1/2  180
 3  2  x  x  2   0   1/2  190
 3  3  1  2  1   1   2/2  110
 3  4  1  2  1   1   2/2  112
 3  5  1  2  1   3   1/1   90
 3  6  1  2  1   3   1/1   92

Simple linear regression analyses are not recommended in data sets including MZ twins. In data sets that include twins variance components analyses can model their increased similarity using the polygenic (g), additive (a) and twin environment (t) components of variance. The Monks model of association should also be valid in data sets including twins.