In Bioconductor/hcaquery: Conversion of single cell experiment output files to relational database

knitr::opts_chunk$set(echo = TRUE)

Setup

library(dplyr)
library(LoomExperiment)
library(hca)
library(hcaquery)

hca package

• explore all available projects

hca::projects()

• explore just those projects for which the organ is "brain" and species is "Homo sapiens"

brain_homo_sapien_filter <- hca::filters(
    genusSpecies = list(is = "Homo sapiens"),
    organ = list(is = c("brain")),
    workflow = list(is = c("optimus_v4.2.2")))

brain_homo_sapien_proj <- hca::projects(filters = brain_homo_sapien_filter)

brain_homo_sapien_proj

• Select a project to look at in greater depth

project_id <- brain_homo_sapien_proj$projectId[3]

• What files are associated with this project

file_filter <- hca::filters(projectId = list(is = project_id))
proj_files <- hca::files(filter = file_filter)
proj_files

• We want loom files for which there is a single donor (taking the two smallest)

additional_file_filters <- hca::filters(projectId = list(is = project_id),
                                        fileFormat = list(is = "loom"),
                                        donorCount = list(is = c(1)))
proj_files_refined <- hca::files(filter = additional_file_filters,
                                 sort = "fileSize",
                                 order = "asc")
proj_files_refined

• download these files

project_file_locations <- proj_files_refined %>%
                            hca::files_download()
project_file_locations

hcaquery package

• let's take the smallest two .loom files from the Optimus workflow across all projects

test_file_filter <- hca::filters(workflow = list(is = c("optimus_v4.2.2")),
                                 fileFormat = list(is = c("loom")))
test_file_tbl <- hca::files(filters = test_file_filter,
                            size = 2, sort = "fileSize", order = "asc")
project_file_locations <- test_file_tbl %>% hca::files_download()

• let's take a look at the data of these files

exp_1 <- LoomExperiment::import(project_file_locations[1],
                                   type = "SingleCellLoomExperiment")
metadata(exp_1)

exp_2 <- LoomExperiment::import(project_file_locations[2],
                                   type = "SingleCellLoomExperiment")
metadata(exp_2) 

• perhaps we could load them into the relational database for better querying

hcaquery::files_to_db(test_file_tbl)

• what tables are available and a description of each

hcaquery::available_tables()

hcaquery::table_description("genes_tbl")
hcaquery::table_description("cells_tbl")
hcaquery::table_description("assays_tbl")
hcaquery::table_description("experiment_overviews")

• query for file with donor species of Homo sapiens

## example of queries on accessible db tables
hca_tbls <- hca_tables()
human_experiments <- hca_tbls$experiments_tbl %>%
    filter(donor_organism_genus_species == "Homo sapiens")

human_experiments

• query for specific genes within a file, returning a SingleCellExperiment object

genes <- c("Gm18956", "Gm37180", "Gm37363")
smat <- hca_file_gene_query(genes = genes,
                            file_ident = "b077b456-6362-59ea-a5b6-230dd8bc6d91")
smat

Bioconductor/hcaquery documentation built on Dec. 17, 2021, 10:56 a.m.