Last updated: 2022-08-30

Checks: 7 0

Knit directory: scATACseq-topics/

This reproducible R Markdown analysis was created with workflowr (version 1.7.0). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

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.

Environment: empty

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.

Seed: set.seed(20200729)

The command set.seed(20200729) 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.

Session information: recorded

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

Cache: none

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

File paths: relative

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

Repository version: e931791

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 e931791. 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:    .DS_Store
    Ignored:    analysis/figure/
    Ignored:    data/.DS_Store
    Ignored:    data/Buenrostro_2018/
    Ignored:    data/Cusanovich_2018/processed_data/Cusanovich_2018_Kidney.RData
    Ignored:    output/Buenrostro_2018/binarized/filtered_peaks/de-buenrostro2018-k=10-noshrink.RData
    Ignored:    output/Buenrostro_2018/binarized/filtered_peaks/fit-Buenrostro2018-binarized-filtered-scd-ex-k=10.rds
    Ignored:    output/Buenrostro_2018/binarized/filtered_peaks/fit-Buenrostro2018-binarized-filtered-scd-ex-k=8.rds
    Ignored:    output/Cusanovich_2018/
    Ignored:    scripts/seq_gene.md.gz

Untracked files:
    Untracked:  analysis/fit-Buenrostro2018-binarized-scd-ex-k=10.rds
    Untracked:  data/Buenrostro_2018_binarized.RData
    Untracked:  output/Buenrostro_2018/binarized/filtered_peaks/Buenrostro_2018_binarized_filtered.RData
    Untracked:  plots/
    Untracked:  scripts/fit-buenrostro-2018-k=8.rds

Unstaged changes:
    Deleted:    analysis/explore_cicero.R

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/explore_cicero.Rmd) and HTML (docs/explore_cicero.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 e931791 Peter Carbonetto 2022-08-30 workflowr::wflow_publish("analysis/explore_cicero.Rmd", verbose = TRUE)
html 69df48d Peter Carbonetto 2022-08-30 Added more text to explore_cicero analysis.
Rmd 22bb834 Peter Carbonetto 2022-08-30 workflowr::wflow_publish("analysis/explore_cicero.Rmd", verbose = TRUE)
Rmd 84c54c8 Peter Carbonetto 2022-08-29 A couple small improvements to the explore_cicero analysis.
Rmd 89de024 Peter Carbonetto 2022-08-29 Adding another plot to the explore_cicero analysis.
html ff33f2d Peter Carbonetto 2022-08-29 Added scatterplot to explore_cicero analysis.
Rmd d1f8485 Peter Carbonetto 2022-08-29 workflowr::wflow_publish("analysis/explore_cicero.Rmd")
Rmd 7f6b8a5 Peter Carbonetto 2022-08-29 Added steps to explore_cicero analysis to load topic model fit and de_analysis results.
Rmd ac066af Peter Carbonetto 2022-08-29 Added some text to the explore_cicero analysis.
html 83686bc Peter Carbonetto 2022-08-29 First build of the explore_cicero workflowr analysis.
Rmd 522f408 Peter Carbonetto 2022-08-29 workflowr::wflow_publish("analysis/explore_cicero.Rmd")
Rmd 39fecb3 Peter Carbonetto 2022-08-29 workflowr::wflow_publish("analysis/index.Rmd")

The aim of this short analysis is to get a better understanding of the Cicero co-accessibility data, and how (and whether) these data can be used to connect chromatin accessibility peaks to genes in order to identify “driving genes” for topics estimated from single-cell ATAC-seq data. As an illustration, here we focus on the Cicero data for a single gene, Slc12a1, that was highlighted in Cusanovich et al, 2018 in connection with the “loop of henle” cell type (see Fig. 5 of that paper, and see also Park et al, 2018).

Load the packages used in the analysis below.

library(fastTopics)
library(ggplot2)
library(cowplot)
library(ashr)

Load the base-pair positions of the genes for the mm9 Mouse Genome Assembly.

load("data/mm9_seq_gene.RData")

Load the Cicero co-accessibility data, including the “gene activity scores”, for gene Slc12a1. (These data were downloaded from the Mouse sci-ATAC-seq Atlas website then prepared using the extract_slc12a1_data.R script.)

load("data/Cusanovich_2018/processed_data/slc12a1_data.RData")
cicero <- transform(cicero,
                    Peak1 = as.character(Peak1),
                    Peak2 = as.character(Peak2))

Load the \(K = 10\) topic model fit, and the results of the DE analysis using this topic model (without the adaptive shrinkage step).

fit <- readRDS(file.path("output/Cusanovich_2018/tissues",
                         "fit-Cusanovich2018-Kidney-scd-ex-k=10.rds"))$fit
fit <- poisson2multinom(fit)
load(file.path("output/Cusanovich_2018/tissues",
           "de-cusanovich2018-kidney-k=10-noshrink.RData"))

From the Structure plots here, topic 8 appears to capture Loop of Henle (LoH) cells, so in the remainder we focus on topic 8.

k <- 8

Get the base-pair positions of the peaks.

feature_names <- rownames(de$postmean)
out           <- strsplit(feature_names,"_")
positions     <- data.frame(chr   = sapply(out,"[[",1),
                            start = sapply(out,"[[",2),
                            end   = sapply(out,"[[",3),
                            name  = feature_names,
                            stringsAsFactors = FALSE)
positions <- transform(positions,
                       start = as.numeric(start),
                       end   = as.numeric(end))

Before examining the co-accessibility data in detail, this first plot confirms that Slc12a1 is highly relevant to topic 8, the LoH topic. It is a simple scatterplot showing the Slc12a1 gene activity score and topic proportion for each cell. The dashed blue line shows the “best fit” line between the two quantities. (Note that the gene activity scores are shown on the log-scale.)

pdat <- data.frame(loading = fit$L[,k],score = log10(1 + scores))
b <- coef(lm(score ~ loading,data = pdat))
ggplot(pdat,aes(x = loading,y = score)) +
  geom_point() +
  geom_abline(intercept = b["(Intercept)"],slope = b["loading"],
              color = "dodgerblue",size = 1,linetype = "dashed") +
  labs(x = "topic proportion",y = "gene activity score") +
  theme_cowplot()

Version Author Date
ff33f2d Peter Carbonetto 2022-08-29

Having confirmed the strong relationship between Slc12a1 and topic 8, let’s examine more closely the de_analysis results for the peaks near Slc12a1. Let’s use a window of 400 kb near the transcribed region of Slc12a1 (124.97–125.06 Mb on chromosome 2).

d <- 4e5
seq_gene <- subset(seq_gene,feature_name == "Slc12a1")
rows <- which(with(positions,chr == paste0("chr",seq_gene$chromosome) &
                   start > seq_gene$chr_start - d &
                   end < seq_gene$chr_stop + d))
peaks <- c(cicero$Peak1,cicero$Peak2)

These next two plots show the p-values (top) and the LFC estimates for topic 8 and for the 111 peaks near Slc12a1. The genes that are identified by Cicero as being relevant to Slc12a1 are highlighted in magenta.

pdat <- data.frame(start      = positions[rows,"start"]/1e5,
                   postmean   = de$postmean[rows,k],
                   z          = de$z[rows,k],
                   lpval      = de$lpval[rows,k],
                   cicero_hit = is.element(positions[rows,"name"],peaks))
p1 <- ggplot(pdat,aes(x = start,y = lpval,color = cicero_hit)) +
  geom_point() +
  scale_color_manual(values = c("royalblue","magenta")) +
  labs(x = "position on chromosome 2 (kb)",
       y = "-log10 p-value") +
  theme_cowplot()
p2 <- ggplot(pdat,aes(x = start,y = postmean,color = cicero_hit)) +
  geom_point() +
  geom_hline(yintercept = 0,color = "black",linetype = "dashed") +
  scale_color_manual(values = c("royalblue","magenta")) +
  labs(x = "position on chromosome 2 (kb)",
       y = "LFC estimate") +
  theme_cowplot()
plot_grid(p1,p2,nrow = 2,ncol = 1)

Version Author Date
69df48d Peter Carbonetto 2022-08-30

It is interesting that most of the Cicero-identified peaks show good evidence for being more accessible in topic 8. But it is also interesting that many other peaks near the gene also show good evidence for being more accessible in topic 8.

To improve the LFC estimates, a simple thing we can do is run ash separately on the subset of peaks that are near Slc12a1. By focussing on this small subset of peaks, ash should recognize that there is a strong signal near the gene, and not shrink the estimates too strongly.

pdat <- transform(pdat,se = postmean/z)
fit <- ash(pdat$postmean,pdat$se,mixcompdist = "normal",method = "shrink")

Indeed, ash preserves the strongest signals, and shrinks the others to zero, or close to zero. It is also interesting that the effects of all the Cicero-identified peaks remain after the adaptive shrinkage step:

pdat$ashmean <- fit$result$PosteriorMean
ggplot(pdat,aes(x = postmean,y = ashmean,color = cicero_hit)) +
  geom_point() +
  geom_abline(intercept = 0,slope = 1,color = "black",linetype = "dashed") +
  scale_color_manual(values = c("royalblue","magenta")) +
  labs(x = "original LFC estimate",y = "ash LFC estimate") +
  theme_cowplot()


sessionInfo()
# R version 3.6.2 (2019-12-12)
# Platform: x86_64-apple-darwin15.6.0 (64-bit)
# Running under: macOS Catalina 10.15.7
# 
# 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] ashr_2.2-54        cowplot_1.1.1      ggplot2_3.3.6      fastTopics_0.6-131
# [5] workflowr_1.7.0   
# 
# loaded via a namespace (and not attached):
#  [1] mcmc_0.9-6         fs_1.5.2           progress_1.2.2     httr_1.4.2        
#  [5] rprojroot_1.3-2    tools_3.6.2        backports_1.1.5    bslib_0.3.1       
#  [9] utf8_1.1.4         R6_2.4.1           irlba_2.3.3        uwot_0.1.10       
# [13] DBI_1.1.0          lazyeval_0.2.2     colorspace_1.4-1   withr_2.5.0       
# [17] tidyselect_1.1.1   prettyunits_1.1.1  processx_3.5.2     compiler_3.6.2    
# [21] git2r_0.29.0       quantreg_5.54      SparseM_1.78       plotly_4.9.2      
# [25] labeling_0.3       sass_0.4.0         scales_1.1.0       SQUAREM_2017.10-1 
# [29] quadprog_1.5-8     callr_3.7.0        pbapply_1.5-1      mixsqp_0.3-46     
# [33] stringr_1.4.0      digest_0.6.23      rmarkdown_2.11     MCMCpack_1.4-5    
# [37] pkgconfig_2.0.3    htmltools_0.5.2    fastmap_1.1.0      invgamma_1.1      
# [41] highr_0.8          htmlwidgets_1.5.1  rlang_0.4.11       rstudioapi_0.13   
# [45] jquerylib_0.1.4    generics_0.0.2     farver_2.0.1       jsonlite_1.7.2    
# [49] dplyr_1.0.7        magrittr_2.0.1     Matrix_1.4-2       Rcpp_1.0.8        
# [53] munsell_0.5.0      fansi_0.4.0        lifecycle_1.0.0    stringi_1.4.3     
# [57] whisker_0.4        yaml_2.2.0         MASS_7.3-51.4      Rtsne_0.15        
# [61] grid_3.6.2         parallel_3.6.2     promises_1.1.0     ggrepel_0.9.1     
# [65] crayon_1.4.1       lattice_0.20-38    hms_1.1.0          knitr_1.37        
# [69] ps_1.6.0           pillar_1.6.2       glue_1.4.2         evaluate_0.14     
# [73] getPass_0.2-2      data.table_1.12.8  RcppParallel_5.1.5 vctrs_0.3.8       
# [77] httpuv_1.5.2       MatrixModels_0.4-1 gtable_0.3.0       purrr_0.3.4       
# [81] tidyr_1.1.3        assertthat_0.2.1   xfun_0.29          coda_0.19-3       
# [85] later_1.0.0        viridisLite_0.3.0  truncnorm_1.0-8    tibble_3.1.3      
# [89] ellipsis_0.3.2