hanziqiang
我上面说的计算Chi-squared的方法不准确,忘掉吧= =
仔细看了一下,heterogeneity statistics,感觉I-squared test 比较符合你的需求
crosstype1为例, 先算 Cochran’s Q, w_i是每个family的样本大小的权值,\(\frac{A}{B}\)
是将7个family的ab合起来算的分离比,\(\frac{A_i}{B_i}\)
是每个family单独的分离比,然后
\(Q=\sum{(\frac{A}{B} - w_i\frac{A_i}{B_i})}^2\)
然后根据Q算I-squared,七个family故 df=7-1 = 6
\(I^2 = 100\% ×(Q - df)/Q\)
解读I2 可以参考
0% to 40%: might not be important
30% to 60%: moderate heterogeneity
50% to 90%: substantial heterogeneity
75% to 100%: considerable heterogeneity
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC192859/