Where’s my gene prototype

Datasets

The following data sets are included in this notebook. Datasets were sub-sampled to a max 200,000 K cells per dataset.

Dataset	Cells	Collection	cellxgene
Individual Single-Cell RNA-seq PBMC Data from Arunachalam et al.	49,139	Aronow	Link
Individual Single-Cell RNA-seq PBMC Data from Schulte-Schrepping et al.	90,957	Aronow	Link
Individual Single-Cell RNA-seq PBMC Data from Lee et al.	43,512	Aronow	Link
Individual Single-Cell RNA-seq PBMC Data from Wilk et al.	41,305	Aronow	Link
Individual Single-Cell RNA-seq PBMC Data from Guo et al.	14,783	Aronow	Link
Azimuth meta-analysis from Adams et al.	312,928	Satija	Link
Azimuth meta-analysis from Delorey et al.	106,043	Satija	Link
Azimuth meta-analysis from Habermann et al.	114,396	Satija	Link
Large-scale single-cell analysis reveals critical immune characteristics of COVID-19 patients	1,462,702	Zemin Zenhg	Link

Read data

This will load all data from the datasets above. Only run this once per session

all_data <- readRDS("all_data.rds")

Parameters

Input the genes and cell types of interest/

Gene set 1 - minimal

#genes <- c("MALAT1", "CD68", "CD79A")
#cell_types <- c("B cell", "T cell", "neutrophil")

Gene set 2 - marker genes

genes <- c("MALAT1", "CD68", "CD79A", "CD8A", "FOXJ1", "GNLY", "CD4", "GNLY", "CD79A", "JCHAIN", "MNDA", "MUC5B", "KTR18")
cell_types <- c("B cell", "T cell", "neutrophil", "natural killer cell", "plasma cell", "macrophage", "ciliated cell")

Gene set 3 - constitutive genes

#genes <- c("LDHA", "PGK1", "ENO1", "SKP1", "TGFB1", "CDKN1C", "PKM", "CCND3", "SKP1", "ZBTB17", "MALAT1", "GAPDH", "ACTB", "RP9", "MT3", "MTR")
#cell_types <- c("B cell", "T cell", "neutrophil", "natural killer cell", "plasma cell", "macrophage", "ciliated cell", "epithelial cell")

Visual prototypes

Any time genes and/or cell types are updated this code has to be run again.

# Summarizing into median, mean, and percent cells
current_data <- get_expression(all_data, genes, cell_types)

Processing dataset  1 / 10 
Processing dataset  2 / 10 
Processing dataset  3 / 10 
Processing dataset  4 / 10 
Processing dataset  5 / 10 
Processing dataset  6 / 10 
Processing dataset  7 / 10 
Processing dataset  8 / 10 
Processing dataset  9 / 10 
Processing dataset  10 / 10 

dot_data <- current_data %>%
  group_by (Gene, cell_type) %>%
  summarise(mean = mean(Expression[Expression!=0]), median = median(Expression[Expression!=0]), percent_cells = sum(Expression != 0) / n()) %>%
  ungroup()

# Appending summary to all data
current_data <- current_data %>%
  dplyr::filter(Expression != 0)

ids <- str_c(current_data$Gene, current_data$cell_type)
mapping <- setNames(dot_data$median, str_c(dot_data$Gene, dot_data$cell_type))
current_data$median <- mapping[ids]

Dot plot: 2-color gradient

plot of chunk sub_chunk_7982

Dot plot: 3-color gradient

plot of chunk sub_chunk_9892

Dot plot: cellxgene colors

plot of chunk sub_chunk_3948

Density plot: 2-color gradient

plot of chunk sub_chunk_3075

Density plot: 3-color gradient

plot of chunk sub_chunk_8272

Density plot: viridis

plot of chunk sub_chunk_8620