In pmartR/pmartRmems:

knitr::opts_chunk$set(echo = FALSE)

library(ALDEx2)
library(pmartRmems)
library(ggplot2)
library(tidyverse)
data("selex")
# str(selex)

conds <- c(rep("NS", 7), rep("S", 7))
# x <- aldex.clr(selex, conds, mc.samples=128, denom="iqlr", verbose=TRUE)
# xt_mean <- aldex.ttest(x, conds, paired.test=TRUE)
# saveRDS(xt_mean, "data-raw/selex_mean.rds")
xt_mean <- readRDS("selex_mean.rds")
# xt_median <- aldex.ttest(x, conds, paired.test=TRUE)
# saveRDS(xt_median, "data-raw/selex_median.rds")
xt_median <- readRDS("selex_median.rds")
pdat <- as.data.frame(cbind(xt_mean, xt_median)) 
names(pdat) <- c("mean_p", "mean_p_adj", "med_p", "med_p_adj")
pdat <- pdat %>% mutate(OTU = rownames(.)) %>% select(OTU, everything())
# plotdat <- pdat %>% gather("method", "p_val", 1:4) %>% mutate(sig = p_val < 0.05)
sig <- 0.05
noadjdat <- pdat %>% select(OTU, mean_p, med_p) %>% 
  mutate(which_sig = ifelse(mean_p < sig & med_p > sig, "mean_p", 
                            ifelse(mean_p > sig & med_p < sig, "med_p", 
                                   ifelse(mean_p < sig & med_p < sig, "both", "neither"))))
adjdat <- pdat %>% select(OTU, mean_p_adj, med_p_adj) %>% 
  mutate(which_sig = ifelse(mean_p_adj < sig & med_p_adj > sig, "mean_p", 
                            ifelse(mean_p_adj > sig & med_p_adj < sig, "med_p", 
                                   ifelse(mean_p_adj < sig & med_p_adj < sig, "both", "neither"))))

theme_set(theme_grey(base_size = 16)) 
ggplot(noadjdat) + geom_point(aes(mean_p, med_p, col = which_sig)) + 
  labs(title = "Unadjusted P-values") + guides(colour = guide_legend(override.aes = list(size=4)))
ggplot(adjdat) + geom_point(aes(mean_p_adj, med_p_adj, col = which_sig)) + 
  labs(title = "Adjusted P-values") + guides(colour = guide_legend(override.aes = list(size=4)))