Merge pull request #101 from shirmath/main

add plotting demo to radEmu vignettes

vignettes/intro_radEmu.Rmd

@@ -158,30 +158,29 @@ ch_fit
 ```
 
 
-The way to access estimated coefficients and confidence intervals from the model is with `ch_fit$coef`. Let's plot our results:
-
-```{r, fig.width= 7, fig.height = 4}
-ch_df <- ch_fit$coef %>%
-  mutate(Genus = (mOTU_name_df %>% 
-           filter(genus_name %in% chosen_genera) %>% 
-           pull(genus_name))[-category_to_rm]) %>%
-  # add genus name to output from emuFit
+
+Now, we can easily visualize our results using the `plot` function!
+So that we only produce the plot, and not the dataframe used to produce the plots,
+we will explicitly extract the plots using the `$` operator to get the `plots` component.
+
+```{r, fig.height = 6, fig.width = 6}
+plot(ch_fit)$plots
+```
+
+In the plot above, it is a bit difficult to read the taxa names on the y-axis. We can create a data frame that maps the full taxa names in our data to simplified labels and plot using those instead, as shown below.
+```{r, fig.height = 6, fig.width = 6}
+#create data frame that has simplified names for taxa
+taxa_names <- data.frame("category" = ch_fit$coef$category) %>%
   mutate(cat_small = stringr::str_remove(paste0("mOTU_", 
                             stringr::str_split(category, 'mOTU_v2_', simplify = TRUE)[, 2]), 
-                            "\\]")) %>%
-  mutate(cat_small = factor(cat_small, levels = cat_small[order(Genus)]))
-  # reorder mOTU categories by genus
-
-ggplot(ch_df) + 
-  geom_point(aes(x = cat_small, y = estimate,color = Genus), size = .5) +
-  geom_errorbar(aes(x = cat_small, ymin = lower, ymax = upper, color = Genus), width = .25) +
-  theme_bw() +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
-  labs(x = "Category",
-       y = "Estimate") + 
-  coord_cartesian(ylim = c(-5,10))
+                            "\\]"))
+
+
+#produce plot with cleaner taxa labels
+plot(ch_fit, taxon_names = taxa_names)$plots
 ```
 
+
 Interestingly, we estimate a meta-mOTU "unknown Eubacterium [meta_mOTU_v2_7116]" assigned to Eubacteria to have a much higher ratio of abundance (comparing CRC group to control) than is typical across the mOTUs we included in this analysis.
 
 The confidence interval for this effect does not include zero -- but (!!!) the kind of confidence interval that is returned by default by emuFit is not extremely reliable when counts are very skewed or sample size is small-to-moderate. 

vignettes/intro_radEmu_with_phyloseq.Rmd

@@ -141,26 +141,12 @@ ch_fit <- emuFit(formula = ~ Group,
                  run_score_tests = FALSE) 
 ```
 
-The way to access estimated coefficients and confidence intervals from the model is with `ch_fit$coef`. Let's plot our results:
-
-```{r, fig.width= 7, fig.height = 4, eval = phy}
-ch_df <- ch_fit$coef %>%
-  mutate(Genus = as.vector(phyloseq::tax_table(wirbel_china)[, 6])) %>%
-  # add genus name to output from emuFit
-  mutate(cat_small = stringr::str_remove(paste0("mOTU_", 
-                            stringr::str_split(category, 'mOTU_v2_', simplify = TRUE)[, 2]), 
-                            "\\]")) %>%
-  mutate(cat_small = factor(cat_small, levels = cat_small[order(Genus)]))
-  # reorder mOTU categories by genus
-
-ggplot(ch_df) + 
-  geom_point(aes(x = cat_small, y = estimate, color = Genus), size = .5) +
-  geom_errorbar(aes(x = cat_small, ymin = lower, ymax = upper, color = Genus), width = .25) +
-  theme_bw() +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
-  labs(x = "Category",
-       y = "Estimate") + 
-  coord_cartesian(ylim = c(-5,10))
+The way to access estimated coefficients and confidence intervals from the model is with `ch_fit$coef`. 
+
+Now, we can easily visualize our results using the `plot.emuFit` function!
+
+```{r, fig.height = 6, fig.width = 6}
+plot(ch_fit)$plots
 ```
 
 If you'd like to see more explanations of the `radEmu` software and additional analyses of this data, check out the vignette "intro_radEmu.Rmd".