Coevolution between PPI partners detected using correlated changes in selective constraint.

aValues of branch-wise selective constraint (ω) allowed in the computation of Spearman’s correlation between these ω values between paired branches for two proteins with a known PPI in humans (Methods).
bValues of ω from the indicated clades were removed before the calculation of the Spearman’s correlation.
cWe required at least 6 common branches between the two orthologous genes trees for the two interacting proteins: the column indicates the number of PPIs meeting this requirement.
dP-value of the hypothesis test that the real PPI pairs had a higher mean Spearman’s correlation than would be expected, given the distribution of correlations seen from 1000 simulations of the same number of pseudo-PPI pairs drawn from non-interacting proteins (Methods).
eMean of the mean correlation seen from 1000 simulations, each consisting of the same number of pseudo-PPIs from c.

a 针对人类中已知存在蛋白质-蛋白质相互作用（Protein-Protein Interaction，PPI）的两个蛋白质，计算其配对分支间的分支特异性选择约束值（ω）的斯皮尔曼相关系数（Spearman’s correlation）时所允许的ω取值范围（方法详见正文）。
b 在计算斯皮尔曼相关系数前，需移除指定演化支对应的ω取值。
c 我们要求两个相互作用蛋白质的直系同源基因树之间至少存在6个共有分支：该列表示满足此要求的蛋白质-蛋白质相互作用对数量。
d 该P值对应假设检验：真实的蛋白质-蛋白质相互作用对的平均斯皮尔曼相关系数，高于从1000次模拟中获取的、取自非相互作用蛋白质的相同数量伪PPI对的相关系数分布所预期的均值（方法详见正文）。
e 1000次模拟所得平均相关系数的均值，每次模拟均采用c项中相同数量的伪蛋白质-蛋白质相互作用对。