Table 2. search.trend performance as assessed by transforming individual clades (either one clade or two clades) within the tree.
One clade | Two clades | |||
---|---|---|---|---|
Tested Individually | Against each other | |||
Simulated phenotype | Opposite sign | Same sign | ||
BM | 0.060 | 0.070 | 0.060 | |
Trend+ | 1.000 | 0.984 | 1.000 | 0.275 |
Trend- | 0.920 | 0.875 | ||
BM | 0.020 | 0.010 | 0.050 | |
Drift+ | 0.882 | 0.935 | 0.956 | 0.148 |
Drift- | 0.857 | 0.882 |
The distribution of Type I and Type II error rates for search.trend assessed on complex phenotypes simulated to exhibit different patterns in different parts of the tree. Two kinds of patterns are simulated, indicating either an increase (indicated with plus symbol), or a decrease (indicated with the minus symbol) in either ‘trend’ or ‘drift’, pertaining to either a single, or two clades within the tree.