Figure 2

Model fit of three different group kinship algorithms. The model fit (vertical axis) is indicated by twice the negative likelihood (the smaller, the better). The grouping was performed with the un-weighted pair group method with arithmetic mean clustering algorithm at different compression levels (horizontal axis), defined as the average number of individuals per group. The phenotypes are (A) body mass index in human, (B) Orthopedic Foundation for Animals score in dog, and flowering time in both (C) maize and (D) Arabidopsis.