Calculate cell type associations using MAGMA — celltype_associations

Has the option of running multiple analyses with a single function. Assumes that you already have all MAGMA GWAS files precomputed. Precomputed MAGMA GWAS files can be downloaded via the import_magma_files function.

celltype_associations_pipeline(
  ctd,
  ctd_levels = seq_len(length(ctd)),
  ctd_name,
  ctd_species = infer_ctd_species(ctd),
  standardise = TRUE,
  magma_dirs,
  run_linear = TRUE,
  run_top10 = TRUE,
  run_conditional = FALSE,
  upstream_kb = 35,
  downstream_kb = 10,
  suffix_linear = "linear",
  suffix_top10 = "top10",
  suffix_condition = "condition",
  controlledAnnotLevel = 1,
  controlTopNcells = 1,
  force_new = FALSE,
  save_dir = tempdir(),
  nThread = 1,
  version = NULL,
  verbose = TRUE
)

Arguments

ctd: Cell type data structure containing specificity_quantiles.
ctd_levels: Which levels of ctd to iterate the enrichment analysis over.
ctd_name: Used in file names
ctd_species: Species name relevant to the CellTypeDataset (ctd). See list_species for all available species. If ctd_species=NULL (default), the ctd species will automatically be inferred using infer_species.
standardise: Whether to run standardise_ctd first. Provides gene ortholog conversion.
magma_dirs: Path to folders containing the pre-computed MAGMA GWAS files (.gsa.rawand .gsa.out). NOTE: Files within these folders must have the same naming scheme as the folders themselves.
run_linear: Run in linear mode.
run_top10: Run in top 10% mode.
run_conditional: Run in conditional mode.
upstream_kb: How many kb upstream of the gene should SNPs be included?
downstream_kb: How many kb downstream of the gene should SNPs be included?
suffix_linear: This will be added to the linear results file name.
suffix_top10: This will be added to the top 10% results file name.
suffix_condition: This will be added to the conditional results file name.
controlledAnnotLevel: Which annotation level should be controlled for.
controlTopNcells: How many of the most significant cell types at that annotation level should be controlled for?
force_new: [Optional] Force new MAGMA analyses even if the pre-existing results files are detected.
save_dir: Folder to save results in (save_dir=NULL to not save any results).
nThread: Number of threads to parallelise analyses across.
version: MAGMA version to use.
verbose: Print messages.

Value

A list containing the results of each selected celltype associations analysis.

Examples

magma_dirs <- MAGMA.Celltyping::import_magma_files(ids = c("ieu-a-298"))
#> Using built-in example files: ieu-a-298.tsv.gz.35UP.10DOWN
#> Returning MAGMA directories.
ctd <- ewceData::ctd()
#> see ?ewceData and browseVignettes('ewceData') for documentation
#> loading from cache

res <- MAGMA.Celltyping::celltype_associations_pipeline(
    ctd = ctd,
    ctd_levels = 1,
    ctd_name = "Zeisel2015",
    ctd_species = "mouse",
    magma_dirs = magma_dirs)
#> Preparing CellTypeDataset.
#> Converting to sparse matrix.
#> Converting to sparse matrix.
#> ieu-a-298.tsv.gz.35UP.10DOWN
#> ======= Calculating celltype associations: linear mode =======
#> Installed MAGMA version: v1.10
#> Skipping MAGMA installation.
#> The desired_version of MAGMA is currently installed: v1.10
#> Using: magma_v1.10
#> Running MAGMA: Linear mode
#> Mapping gene symbols in specificity_quantiles matrix to entrez IDs.
#> Reading enrichment results file into R.
#> ======= Calculating celltype associations: top10% mode =======
#> Installed MAGMA version: v1.10
#> Skipping MAGMA installation.
#> The desired_version of MAGMA is currently installed: v1.10
#> Using: magma_v1.10
#> Running MAGMA: Top 10% mode
#> Mapping gene symbols in specificity_deciles matrix to entrez IDs.
#> Constructing top10% gene marker sets for 8 cell-types.
#> Reading enrichment results file into R.
#> Merging linear and top10% results
#> Saving results ==> /tmp/RtmpUbtMhH/Zeisel2015/MAGMA_celltyping.Zeisel2015.rds