Last updated: 2020-08-19

Checks: 7 0

Knit directory: single-cell-topics/analysis/

This reproducible R Markdown analysis was created with workflowr (version 1.6.2.9000). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.


Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

The command set.seed(1) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version c7b77ee. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    data/droplet.RData
    Ignored:    data/pbmc_68k.RData
    Ignored:    data/pbmc_purified.RData
    Ignored:    data/pulseseq.RData
    Ignored:    output/droplet/fits-droplet.RData
    Ignored:    output/droplet/rds/
    Ignored:    output/pbmc-68k/fits-pbmc-68k.RData
    Ignored:    output/pbmc-68k/rds/
    Ignored:    output/pbmc-purified/fits-pbmc-purified.RData
    Ignored:    output/pbmc-purified/rds/
    Ignored:    output/pulseseq/fits-pulseseq.RData
    Ignored:    output/pulseseq/rds/

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.


These are the previous versions of the repository in which changes were made to the R Markdown (analysis/plots_tracheal_epithelium.Rmd) and HTML (docs/plots_tracheal_epithelium.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File Version Author Date Message
Rmd c7b77ee Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
Rmd 70a4a60 Peter Carbonetto 2020-08-19 Added note to plots_tracheal_epithelium.Rmd.
html fb21b3b Peter Carbonetto 2020-08-19 Added very initial Structure plots to plots_tracheal_epithelium analysis.
Rmd d35cb03 Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html 368a74a Peter Carbonetto 2020-08-19 Added some text to plots_tracheal_epithelium analysis.
Rmd 223406b Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html aca46cc Peter Carbonetto 2020-08-19 Added manual clustering of droplet samples based on PCs.
Rmd 38f811b Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
Rmd 343747e Peter Carbonetto 2020-08-19 Small edit to figure dimensions.
html 5a35bbd Peter Carbonetto 2020-08-19 Added labeled PCA plot; adjusted plot dimensions in
Rmd fb91075 Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html 8b9b528 Peter Carbonetto 2020-08-19 Added more PCA plots to plots_tracheal_epithelium analysis.
Rmd ee7cbf1 Peter Carbonetto 2020-08-19 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html c517ea2 Peter Carbonetto 2020-08-18 Small fix to one of the PCA plots in plots_tracheal_epithelium.
Rmd 8f5c210 Peter Carbonetto 2020-08-18 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html 01afbd2 Peter Carbonetto 2020-08-18 Added some PC plots to the plots_tracheal_epithelium analysis.
Rmd f1c7d02 Peter Carbonetto 2020-08-18 wflow_publish(“plots_tracheal_epithelium.Rmd”)
html 0a04fc1 Peter Carbonetto 2020-08-18 Added abundance plots to plots_tracheal_epithelium analysis.
Rmd f914f7e Peter Carbonetto 2020-08-18 wflow_publish(“plots_tracheal_epithelium.Rmd”)
Rmd 61917ad Peter Carbonetto 2020-08-18 Working on new analysis, plots_tracheal_epithelium.Rmd.

TO DO: Add introductory text here.

Load the packages used in the analysis below.

# library(fastTopics)
devtools::load_all("~/git/fastTopics")
library(ggplot2)
library(cowplot)
source("../code/plots.R")

Load the sample annotations. (The count data are no longer needed at this stage of the analysis.)

load("../data/droplet.RData")
samples_droplet <- samples
load("../data/pulseseq.RData")
samples_pulseseq <- samples
rm(samples,counts)

Load the \(k = 7\) Poisson NMF model fit for the droplet data, and the \(k = 11\) Poisson NMF fit for the pulse-seq data.

fit_droplet <- readRDS("../output/droplet/rds/fit-droplet-scd-ex-k=7.rds")$fit
fit_pulseseq <- readRDS("../output/pulseseq/rds/fit-pulseseq-scd-ex-k=11.rds")$fit

The Montoro et al (2018) article mentions that some epithelial cell types are abundant, whereas others are very rare. The topics in the droplet data reflect this:

p1 <- create_abundance_plot(fit_droplet)
print(p1)

Version Author Date
0a04fc1 Peter Carbonetto 2020-08-18

The first topic—which is not visible in the bar chart—is indeed very rare; only 43 out of 7,193 samples have a greater than 10% contribution from this topic.

sum(poisson2multinom(fit_droplet)$L[,1] > 0.1)
# [1] 43

Likewise, we also pick up rare and abundant topics in the pulse-seq data:

p2 <- create_abundance_plot(fit_pulseseq)
print(p2)

Version Author Date
0a04fc1 Peter Carbonetto 2020-08-18

In this next part of the analysis, we perform PCA on the estimated topic proportions to explore structure in the data as inferred by the topic model. Typically a nonlinear embedding method such as t-SNE or UMAP is used to visualize the structure, but the disadvantage such methods is that it can often be difficult to get the (many) tuning parameters right, and they are sometimes very slow for large data sets; by contrast, PCA has no tuning parameters.

These three scatterplots show the droplet samples (the topic proportions) projected onto 5 out of the 7 PCs. (PCs 3 and 7 do not reveal any additional structure, so are not shown here.)

p3 <- basic_pca_plot(fit_droplet,c("PC1","PC2"))
p4 <- basic_pca_plot(fit_droplet,c("PC4","PC5"))
p5 <- basic_pca_plot(fit_droplet,c("PC5","PC6"))
plot_grid(p3,p4,p5,nrow = 1)

Version Author Date
8b9b528 Peter Carbonetto 2020-08-19
01afbd2 Peter Carbonetto 2020-08-18

Distinct clusters show up in the PC1 vs. PC2 plot, as well as in the PC5 vs. PC6 plot, whereas the structure in PC4 vs. PC5 is very much continuously varying.

Let’s look more closely at the topics that show variation in the first two PCs:

p6 <- pca_plot(poisson2multinom(fit_droplet),pcs = 1:2,k = c(2,5,6),
               plot_grid_call = function (plots)
                 do.call(plot_grid,c(lapply(plots,
                   function (x) x + guides(fill = "none")),list(nrow = 1))))
print(p6)

Version Author Date
5a35bbd Peter Carbonetto 2020-08-19
8b9b528 Peter Carbonetto 2020-08-19

The bulk of the samples lie on a continuous gradient between topics 2 and 5. There is a smaller cluster at the bottom of this plot, with high contributions from topic 6.

Next, we look closely at PCs 4 and 5:

p7 <- pca_plot(poisson2multinom(fit_droplet),pcs = 4:5,k = c(3,4,5,7),
               plot_grid_call = function (plots)
                 do.call(plot_grid,c(lapply(plots,
                   function (x) x + guides(fill = "none")),list(nrow = 1))))
print(p7)

Version Author Date
fb21b3b Peter Carbonetto 2020-08-19
5a35bbd Peter Carbonetto 2020-08-19
8b9b528 Peter Carbonetto 2020-08-19

Along these PCs, we see that topics 3, 4, 5 and 7 exist in many combinations, with no apparent distinct populations.

Topic 1 captures a very small discrete population of cells:

p8 <- pca_plot(poisson2multinom(fit_droplet),pcs = 5:6,k = 1) +
        guides(fill = "none")
print(p8)

Version Author Date
8b9b528 Peter Carbonetto 2020-08-19

In summary, topics 1 and 6 pick up discrete “cell types”, whereas the other topics characterize more continuous variation in gene expression, perhaps cell types along a continuous trajectory of development. There are some other discrete clusters that seem to be composed of distinct combinations of topics that we will neeed to examine more closely.

Four suggested by PCA analysis, plus a “background” cluster (\(c_0\)):

droplet_cluster_colors <- c("gray","darkblue","gold","yellowgreen","firebrick")
pca_droplet <- prcomp(poisson2multinom(fit_droplet)$L)$x
n <- nrow(pca_droplet)
x <- rep("c0",n)
pc1 <- pca_droplet[,"PC1"]
pc2 <- pca_droplet[,"PC2"]
pc6 <- pca_droplet[,"PC6"]
x[pc2 > -0.1] <- "c1"
x[pc6 < -0.04] <- "c2"
x[(pc1 - 0)^2 + (pc2 + 0.75)^2 < 0.09] <- "c3"
x[(pc1 - 0.5)^2 + (pc2 + 0.9)^2 < 0.04] <- "c4"
samples_droplet$cluster <- factor(x)
p9 <- pca_plot_with_labels(fit_droplet,c("PC1","PC2"),samples_droplet$cluster,
                           droplet_cluster_colors) + labs(fill = "cluster")
p10 <- pca_plot_with_labels(fit_droplet,c("PC4","PC5"),samples_droplet$cluster,
                            droplet_cluster_colors) + labs(fill = "cluster")
p11 <- pca_plot_with_labels(fit_droplet,c("PC5","PC6"),samples_droplet$cluster,
                            droplet_cluster_colors) + labs(fill = "cluster")
plot_grid(p9,p10,p11)

Version Author Date
fb21b3b Peter Carbonetto 2020-08-19
368a74a Peter Carbonetto 2020-08-19
aca46cc Peter Carbonetto 2020-08-19

The vast majority of the droplet samples are in the \(c_1\) cluster:

table(samples_droplet$cluster)
# 
#   c0   c1   c2   c3   c4 
#   73 6533   50  162  375

This suggests separate treatment of the \(c_1\) cluster.

TO DO: Add more text here.

set.seed(1)
droplet_topic_colors <- c("gold","royalblue","turquoise","lightgreen",
                          "forestgreen","firebrick","olivedrab")
topic_ordering <- c(1,5,6,2,3,4,7)
fit_droplet_rare <- select(poisson2multinom(fit_droplet),
                           loadings = which(samples_droplet$cluster != "c1" &
                                            samples_droplet$cluster != "c4"))
tsne_droplet_rare <- tsne_from_topics(fit_droplet_rare,dims = 1,perplexity = 30,
                                      verbose = FALSE)
p12 <- structure_plot(fit_droplet_rare,rows = order(tsne_droplet_rare$Y),
                      topics = topic_ordering,
                      colors = droplet_topic_colors[topic_ordering])
print(p12)

Version Author Date
fb21b3b Peter Carbonetto 2020-08-19
temp <- select(poisson2multinom(fit_droplet),
               loadings = which(samples_droplet$tissue == "Ionocyte"))
structure_plot(temp,rows = order(temp$L[,3]),
                      topics = topic_ordering,
                      colors = droplet_topic_colors[topic_ordering])

temp2 <- select(poisson2multinom(fit_droplet),
                loadings = which(samples_droplet$tissue == "Tuft"))
structure_plot(temp2,perplexity = 30,
               topics = topic_ordering,
               colors = droplet_topic_colors[topic_ordering])

temp3 <- select(poisson2multinom(fit_droplet),
                loadings = which(samples_droplet$tissue == "Neuroendocrine"))
structure_plot(temp3,perplexity = 30,
               topics = topic_ordering,
               colors = droplet_topic_colors[topic_ordering])

# Read the 158 x 7 data matrix successfully!
# OpenMP is working. 1 threads.
# Using no_dims = 1, perplexity = 30.000000, and theta = 0.100000
# Computing input similarities...
# Building tree...
# Done in 0.01 seconds (sparsity = 0.752924)!
# Learning embedding...
# Iteration 50: error is 46.264071 (50 iterations in 0.01 seconds)
# Iteration 100: error is 45.443002 (50 iterations in 0.01 seconds)
# Iteration 150: error is 44.347590 (50 iterations in 0.01 seconds)
# Iteration 200: error is 45.517446 (50 iterations in 0.01 seconds)
# Iteration 250: error is 46.081133 (50 iterations in 0.01 seconds)
# Iteration 300: error is 0.338211 (50 iterations in 0.02 seconds)
# Iteration 350: error is 0.262554 (50 iterations in 0.01 seconds)
# Iteration 400: error is 0.262338 (50 iterations in 0.01 seconds)
# Iteration 450: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 500: error is 0.262333 (50 iterations in 0.02 seconds)
# Iteration 550: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 600: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 650: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 700: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 750: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 800: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 850: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 900: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 950: error is 0.262333 (50 iterations in 0.01 seconds)
# Iteration 1000: error is 0.262333 (50 iterations in 0.01 seconds)
# Fitting performed in 0.25 seconds.
# Read the 96 x 7 data matrix successfully!
# OpenMP is working. 1 threads.
# Using no_dims = 1, perplexity = 30.000000, and theta = 0.100000
# Computing input similarities...
# Building tree...
# Done in 0.01 seconds (sparsity = 0.983941)!
# Learning embedding...
# Iteration 50: error is 51.369311 (50 iterations in 0.01 seconds)
# Iteration 100: error is 49.505412 (50 iterations in 0.01 seconds)
# Iteration 150: error is 49.533513 (50 iterations in 0.01 seconds)
# Iteration 200: error is 45.739453 (50 iterations in 0.01 seconds)
# Iteration 250: error is 48.708066 (50 iterations in 0.01 seconds)
# Iteration 300: error is 1.416143 (50 iterations in 0.01 seconds)
# Iteration 350: error is 0.382373 (50 iterations in 0.01 seconds)
# Iteration 400: error is 0.312025 (50 iterations in 0.01 seconds)
# Iteration 450: error is 0.312065 (50 iterations in 0.01 seconds)
# Iteration 500: error is 0.312066 (50 iterations in 0.01 seconds)
# Iteration 550: error is 0.312066 (50 iterations in 0.01 seconds)
# Iteration 600: error is 0.312065 (50 iterations in 0.01 seconds)
# Iteration 650: error is 0.312065 (50 iterations in 0.01 seconds)
# Iteration 700: error is 0.312067 (50 iterations in 0.01 seconds)
# Iteration 750: error is 0.312067 (50 iterations in 0.01 seconds)
# Iteration 800: error is 0.312066 (50 iterations in 0.01 seconds)
# Iteration 850: error is 0.312067 (50 iterations in 0.01 seconds)
# Iteration 900: error is 0.312065 (50 iterations in 0.01 seconds)
# Iteration 950: error is 0.312065 (50 iterations in 0.01 seconds)
# Iteration 1000: error is 0.312067 (50 iterations in 0.01 seconds)
# Fitting performed in 0.12 seconds.
set.seed(1)
fit_droplet_abundant <-
  select(poisson2multinom(fit_droplet),
                          loadings = which(samples_droplet$cluster == "c1" |
                                           samples_droplet$cluster == "c4"))
tsne_droplet_abundant <- tsne_from_topics(fit_droplet_abundant,dims = 1,n = 1000,
                                          perplexity = 100,verbose = FALSE)
Y   <- tsne_droplet_abundant$Y
p13 <- structure_plot(fit_droplet_abundant,
                      rows   = rownames(Y)[order(Y)],
                      topics = topic_ordering,
                      colors = droplet_topic_colors[topic_ordering])
print(p13)

It is helpful to compare these results with clustering reported in the Montoro et al (2018) paper. To make this comparison, we layer the 7 clusters on top of these PCs:

droplet_celltype_colors <-
  c("royalblue",      # Basal
    "firebrick",      # Ciliated
    "forestgreen",    # Club
    "gold",           # Goblet
    "darkmagenta",    # Ionocyte
    "darkorange",     # Neuroendocrine
    "lightsteelblue") # Tuft
p9 <- pca_plot_with_labels(fit_droplet,c("PC1","PC2"),samples_droplet$tissue,
                           droplet_celltype_colors) + labs(fill = "celltype")
p10 <- pca_plot_with_labels(fit_droplet,c("PC4","PC5"),samples_droplet$tissue,
                            droplet_celltype_colors) + labs(fill = "celltype")
p11 <- pca_plot_with_labels(fit_droplet,c("PC5","PC6"),samples_droplet$tissue,
                            droplet_celltype_colors) + labs(fill = "celltype")
plot_grid(p9,p10,p11)

Version Author Date
aca46cc Peter Carbonetto 2020-08-19
5a35bbd Peter Carbonetto 2020-08-19
c517ea2 Peter Carbonetto 2020-08-18
01afbd2 Peter Carbonetto 2020-08-18

TO DO: Add Structure plot9s) to compare Montoro et al (2018) clustering.

And likewise in the pulse-seq data:

p5 <- basic_pca_plot(fit_pulseseq,c("PC3","PC4"))
p6 <- basic_pca_plot(fit_pulseseq,c("PC5","PC6"))
plot_grid(p5,p6)
ggplot(cbind(pca$x,data.frame(k2 = fit$L[,2])),
       aes(x = PC1,y = PC2,fill = k2)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(pca$x,data.frame(k5 = fit$L[,5])),
       aes(x = PC1,y = PC2,fill = k5)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(pca$x,data.frame(k6 = fit$L[,6])),
       aes(x = PC1,y = PC2,fill = k6)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(pca$x,data.frame(k1 = fit$L[,1])),
       aes(x = PC5,y = PC6,fill = k1)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(pca$x,data.frame(k3 = fit$L[,3])),
       aes(x = PC1,y = PC2,fill = k3)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(pca$x,data.frame(k4 = fit$L[,4])),
       aes(x = PC1,y = PC2,fill = k4)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_gradient2(low = "deepskyblue",mid = "gold",high = "orangered",
                       midpoint = 0.5) +
  theme_cowplot(font_size = 12)
ggplot(cbind(samples_droplet,pca$x),aes(x = PC1,y = PC2,fill = tissue)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_manual(values = clrs) +
  theme_cowplot(font_size = 12)
ggplot(cbind(samples_droplet,pca$x),aes(x = PC5,y = PC6,fill = tissue)) +
  geom_point(shape = 21,color = "white",,size = 2) +
  scale_fill_manual(values = clrs) +
  theme_cowplot(font_size = 12)
clrs <- c("royalblue",      # basal
          "firebrick",      # ciliated
          "forestgreen",    # club
          "gold",           # goblet
          "darkmagenta",    # ionocyte
          "darkorange",     # neuroendocrine
          "tomato",         # proliferating
          "darkgray")       # tuft
ggplot(cbind(samples_droplet,pca$x),aes(x = PC1,y = PC2,fill = tissue)) +
  geom_point(shape = 21,color = "white",size = 2) +
  scale_fill_manual(values = clrs) +
  theme_cowplot(font_size = 12)
ggplot(cbind(samples_droplet,pca$x),aes(x = PC5,y = PC6,fill = tissue)) +
  geom_point(shape = 21,color = "white",,size = 2) +
  scale_fill_manual(values = clrs) +
  theme_cowplot(font_size = 12)

sessionInfo()
# R version 3.6.2 (2019-12-12)
# Platform: x86_64-apple-darwin15.6.0 (64-bit)
# Running under: macOS Catalina 10.15.5
# 
# Matrix products: default
# BLAS:   /Library/Frameworks/R.framework/Versions/3.6/Resources/lib/libRblas.0.dylib
# LAPACK: /Library/Frameworks/R.framework/Versions/3.6/Resources/lib/libRlapack.dylib
# 
# locale:
# [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
# 
# attached base packages:
# [1] stats     graphics  grDevices utils     datasets  methods   base     
# 
# other attached packages:
# [1] cowplot_1.0.0      ggplot2_3.3.0      fastTopics_0.3-161 testthat_2.3.1    
# 
# loaded via a namespace (and not attached):
#  [1] httr_1.4.1           tidyr_1.0.0          pkgload_1.0.2       
#  [4] viridisLite_0.3.0    jsonlite_1.6         RcppParallel_4.4.4  
#  [7] assertthat_0.2.1     yaml_2.2.0           remotes_2.1.0       
# [10] progress_1.2.2       ggrepel_0.9.0        sessioninfo_1.1.1   
# [13] pillar_1.4.3         backports_1.1.5      lattice_0.20-38     
# [16] quantreg_5.54        glue_1.3.1           quadprog_1.5-8      
# [19] digest_0.6.23        promises_1.1.0       colorspace_1.4-1    
# [22] htmltools_0.4.0      httpuv_1.5.2         Matrix_1.2-18       
# [25] pkgconfig_2.0.3      devtools_2.2.1       SparseM_1.78        
# [28] purrr_0.3.3          scales_1.1.0         processx_3.4.1      
# [31] whisker_0.4          later_1.0.0          Rtsne_0.15          
# [34] MatrixModels_0.4-1   git2r_0.26.1         tibble_2.1.3        
# [37] farver_2.0.1         usethis_1.6.0        ellipsis_0.3.0      
# [40] withr_2.1.2          lazyeval_0.2.2       cli_2.0.0           
# [43] magrittr_1.5         crayon_1.3.4         memoise_1.1.0       
# [46] mcmc_0.9-6           evaluate_0.14        ps_1.3.0            
# [49] fs_1.3.1             fansi_0.4.0          MASS_7.3-51.4       
# [52] pkgbuild_1.0.6       tools_3.6.2          data.table_1.12.8   
# [55] prettyunits_1.1.1    hms_0.5.2            lifecycle_0.1.0     
# [58] stringr_1.4.0        MCMCpack_1.4-5       plotly_4.9.2        
# [61] munsell_0.5.0        irlba_2.3.3          callr_3.4.0         
# [64] compiler_3.6.2       rlang_0.4.5          grid_3.6.2          
# [67] rstudioapi_0.10      htmlwidgets_1.5.1    labeling_0.3        
# [70] rmarkdown_2.3        gtable_0.3.0         R6_2.4.1            
# [73] knitr_1.26           dplyr_0.8.3          zeallot_0.1.0       
# [76] workflowr_1.6.2.9000 rprojroot_1.3-2      desc_1.2.0          
# [79] stringi_1.4.3        Rcpp_1.0.3           vctrs_0.2.1         
# [82] tidyselect_0.2.5     xfun_0.11            coda_0.19-3