Script used during Human Cell Atlas presentation, 2021-10-06

hca2021bioc.R

##
## Preliminiaries: install infrasctructure and required packages
##

## We've got to be using a least R version 4.1
my_R_version <-
    numeric_version(paste(R.Version()[c("major", "minor")], collapse = "."))
stopifnot(my_R_version >= "4.1")


## install pre-requisites for package installation
prerequisites <- c("remotes", "BiocManager", "futile.logger")
prerequisites <-
    prerequisites[!prerequisites %in% rownames(installed.packages())]
install.packages(prerequisites, repos = "https://cran.r-project.org")

## install the 'devel' version of the hca package; may be prompted to
## update out-of-date packages
BiocManager::install("Bioconductor/hca")

pkgs <- "LoomExperiment"
pkgs <- pkgs[!pkgs %in% rownames(installed.packages())]
BiocManager::install(pkgs)

##
## The HCA Data Portal
##

suppressPackageStartupMessages({
    library(hca)
    library(dplyr)
})

## all projects
projects <- projects()
projects

## filter on the `projects` object -- title of interest
projects |>
    filter(grepl("development of the blood", projectTitle)) |>
    glimpse()

## filter on 'facets' of the data: what facets are avaiable?
projects_facets()

projects_facets() |>
    filter(endsWith(facet, "Organ"))

## What values values are available within the 'specimenOrgan' facet?
projects_facets("specimenOrgan")

## find all studies of the liver
filter <- filters(
    specimenOrgan = list(is = "liver")
)
projects(filter)

## I want more information about these studies!
tbl <- projects(filter, as = "tibble_expanded")
names(tbl)

## I want as many donors as possible
tbl |>
    select(projectId, projects.projectTitle, donorOrganisms.donorCount)

tbl |>
    filter(donorOrganisms.donorCount > 5)


##
## Loom files
##

## discovery and retrieval

## one experiment, loom files from the DCP/2 analysis pipeline,
## including post-processing (combine across samples; 'sparse'
## representation) step
filter <- filters(
    projectId = list(is = "4d6f6c96-2a83-43d8-8fe1-0f53bffd4674"),
    fileSource = list(is = "DCP/2 Analysis"),
    fileFormat = list(is = "loom"),
    workflow = list(is = "optimus_post_processing_v1.0.0")
)

loom_file_path <-
    files(filter) |> # find files matching filter
    files_download() # download (and cache) to local disk

## import into R

suppressPackageStartupMessages({
    library(LoomExperiment)
})
loom <- LoomExperiment::import(loom_file_path)
loom

## add biological annotations

loom_annotated <- optimus_loom_annotation(loom)
table(loom_annotated$input_id, loom_annotated$donor_organism.sex)