在绘制figure时，有连续性变量赋值时常常要用colorbar表示值的大小，如果有特殊的界值（如Pvalue<0.05, OR>1），需要将中间值单独设置个颜色作为区分。以我的自建数据集为例。

当不设置区分时

ggplot(data = dat1,aes(x = Dictionary.element,y = log(OR),color=-log(FDR))) +
  geom_point() +
  theme_bw()+
  geom_hline(yintercept = 1,color="red",linetype="dotted")+
  labs(x="Dictionary element",y="OR (95%CI)-log transformed")

此时无法判断哪个点为有意义。scales包中的scale_color_gradient2可以解决这个问题。 

ggplot(data = dat1,aes(x = Dictionary.element,y = log(OR),color=-log(FDR))) +
  geom_point() +
  theme_bw()+
  geom_hline(yintercept = 1,color="red",linetype="dotted")+
  labs(x="Dictionary element",y="OR (95%CI)-log transformed")+
  scale_color_gradient2(midpoint = 3, mid = "white", 
                        high = muted("red"), low = muted("blue"), 
                        guide = "colourbar", space = "Lab")

 2)方法二，scale_fill_gradientn

ggplot(data = dat1,aes(x = group,y = log(OR),color=-log(FDR))) +
    geom_point() +
    theme_bw()+
    geom_hline(yintercept = 1,color="red",linetype="dotted")+
    labs(x="Dictionary element",y="OR (95%CI)-log- 
    transformed")+scale_color_gradientn(colours = c("blue","white","red"), values = 
    rescale(c(1,3,15)),guide = "colorbar", limits=c(1,15))

 想要设置对称时，可以修改limit和rescale范围，如

scale_color_gradientn(colours = c("blue","white","red"), values = rescale(c(-9,3,15)),guide = "colorbar", limits=c(-9,15))

 此时为以3位对称的colorbar

 Reference:

r - asymmetric color distribution in scale_gradient2? - Stack Overflow