fix ResetModuleNames

    seurat_obj <- SetEnrichrTable(seurat_obj, enrich_table, wgcna_name)

  }

  # update ROC info:

  # THIS DOES NOT UPDATE THE ROC OBJECTS THEMSELVES!!!

  roc_data <- GetROCData(seurat_obj, wgcna_name)

  if(!is.null(roc_data)){

    roc_data$roc$module <- factor(

      new_mod_df[match(roc_data$roc$module, new_mod_df$old),'new'],

      levels = as.character(new_mod_df$new)

    )

    roc_data$conf$module <- factor(

      new_mod_df[match(roc_data$conf$module, new_mod_df$old),'new'],

      levels = as.character(new_mod_df$new)

    )

    seurat_obj <- SetROCData(seurat_obj, roc_data, wgcna_name)

  }

  # update motif overlap

  overlap_df <- GetMotifOverlap(seurat_obj, wgcna_name)

  if(!is.null(overlap_df)){

}

#' ROCCurves

#'

#' Makes barplots from Enrichr data

#'

#' @param seurat_obj A Seurat object

#' @param dbs List of EnrichR databases

#' @param max_genes Max number of genes to include per module, ranked by kME.

#' @param wgcna_name The name of the hdWGCNA experiment in the seurat_obj@misc slot

#' @keywords scRNA-seq

#' @export

#' @examples

#' ROCCurves

ROCCurves <- function(

  seurat_obj,

  roc_df=NULL,

  conf_df=NULL,

  wgcna_name=NULL

){

  if(is.null(wgcna_name)){wgcna_name <- seurat_obj@misc$active_wgcna}

  # get Modules

  modules <- GetModules(seurat_obj)

  mods <- levels(modules$module)

  mods <- mods[mods != 'grey']

  # get module colors:

  mod_colors <- modules %>% subset(module %in% mods) %>%

    select(c(module, color)) %>%

    distinct %>%

    arrange(module) %>% .$color

  # get the ROC info from seurat obj:

  if(is.null(roc_df) | is.null(conf_df)){

    roc_df <- GetROCData(seurat_obj, wgcna_name)$roc

    conf_df <- GetROCData(seurat_obj, wgcna_name)$conf

  }

  # plot the ROC curve

  roc_df <- roc_df %>% group_by(module) %>% arrange(sensitivity)

  conf_df <- conf_df %>% group_by(module) %>% arrange(sensitivity)

  auc_df <- distinct(roc_df[,c('module', 'auc')])

  # set factor levels for modules:

  roc_df$module <- factor(as.character(roc_df$module), levels=mods)

  conf_df$module <- factor(as.character(conf_df$module), levels=mods)

  auc_df$module <- factor(as.character(auc_df$module), levels=mods)

  p <- roc_df %>% ggplot(

    aes(x=specificity, y=sensitivity, color=module, fill=module),

  ) +

    geom_line() +

    geom_ribbon(

      data=conf_df,

      aes(x = sensitivity, ymin=lo, ymax=hi, fill=module),

      inherit.aes=FALSE, alpha=0.4

    ) +

    scale_color_manual(values = unlist(mod_colors)) +

    scale_fill_manual(values = unlist(mod_colors)) +

    scale_x_continuous(breaks = c(0, 0.5, 1), labels=c("0", "0.5", "1")) +

    scale_y_continuous(breaks = c(0, 0.5, 1), labels=c("0", "0.5", "1")) +

    xlab("1 - Specificity (FPR)") + ylab("Sensitivity (TPR)") +

    geom_text(

      data = auc_df,

      aes(color=module),

      x=0.75, y=0.1, label=paste0("AUC: ", format(auc_df$auc, digits=2)),

      inherit.aes=FALSE, size=4, color='black'

    )

  p

}

#' Displays the top n TFs in a set of modules as a bar plot

#'

#' @param seurat_obj A Seurat object

```

<img src="figures/projection/module_preservation_all.png" width="700" height="700">

<!--

## Binary classification analysis

Explain the binary classification analysis.

### Transfer cell states from reference to query

```{r eval=FALSE}

# add SCANVI predicted labels from Zhou et al:

predicted_labels <- read.csv(paste0('../data/Swarup_2021_predicted_obs.csv'))

predicted_labels <- predicted_labels %>% subset(X %in% colnames(seurat_query))

seurat_query$predictions <- predicted_labels$predictions

# setup cell cluster labels for ROC computation

seurat_ref$roc_group <- factor(

  as.character(seurat_ref$annotation),

  levels = levels(seurat_ref$annotation)

)

seurat_query$roc_group <- factor(

  as.character(seurat_query$predictions),

  levels = levels(seurat_ref$roc_group)

)

```

Now we visualize the transferred cell states on the UMAP:

<details> <summary> Code </summary>

```{r eval=FALSE}

p1 <- DimPlot(seurat_ref, group.by='roc_group', label=TRUE, repel=TRUE) +

   umap_theme() +

   ggtitle('Zhou') +

   NoLegend()

p2 <- DimPlot(seurat_query, group.by='roc_group', label=TRUE, repel=TRUE) +

   umap_theme() +

   ggtitle('Morabito & Miyoshi') +

   NoLegend()

p1 | p2

```

</details>

<img src="figures/projection/umap_transfer_labels.png" width="600" height="600">

### Compute ROC

```{r eval=FALSE}

# compute ROC

seurat_ref <- ComputeROC(

  seurat_ref,

  seurat_test = seurat_query,

  group.by = 'roc_group'

)

# head the ROC table:

roc_data <- GetROCData(seurat_obj)

head(roc_data$roc)

```

<details> <summary> Output </summary>

```

specificity sensitivity module   color       auc

1   1.0000000           1 INH-M1 darkred 0.9948718

2   0.9333333           1 INH-M1 darkred 0.9948718

3   0.8666667           1 INH-M1 darkred 0.9948718

4   0.8000000           1 INH-M1 darkred 0.9948718

5   0.7333333           1 INH-M1 darkred 0.9948718

6   0.6666667           1 INH-M1 darkred 0.9948718

```

</details>

### Plot ROC curves

```{r eval=FALSE}

# plot the ROC Curves

p <- ROCCurves(seurat_obj=seurat_obj)

# plot the ROC curves faceted by module

p + facet_wrap(~module, ncol=6) + NoLegend()

```

<img src="figures/projection/Zhou_ROC.png" width="700" height="700"> -->