Package 'mrpipeline'

Description

Calculates proportion of variance explained. From S1 Text

Usage

.calc_pve(b, maf, se, n)

Arguments

Value

Vector or number, proportion of variance explained

Description

Helper function to extract colocalisation regions for when one dataset comes from a local file and another from OpenGWAS.

Usage

.cdat_from_mixed(f1, f2, chrpos, verbose = TRUE)

Arguments

Value

list of coloc-ready data

See Also

[gwasglue::gwasvcf_to_coloc()], [gwasglue::ieugwasr_to_coloc()]

Description

Splits vector into chunks

Usage

.chunk(x, n)

Arguments

Value

list of chunks

Description

Sub-function for the colocalisation analyses

Usage

.coloc_sub(
  dat1,
  dat2,
  min_snps = 100,
  p1 = 1e-04,
  p2 = 1e-04,
  p12 = 1e-05,
  susie = FALSE,
  bfile = NULL,
  plink = NULL,
  verbose = TRUE
)

Arguments

Value

Results data.frame

Description

Sub function to run SuSiE and coloc

Usage

.coloc_susie_sub(d1, d2, bfile = NULL, plink = NULL, verbose = TRUE, ...)

Arguments

Value

Results data.frame

Description

Lookup ENSGs using the Drug Genome Interaction DB API.

Usage

.dgidb_linkage(ensgs)

Arguments

Value

data.frame of results

Description

Attempts to convert ENSG IDs to gene names (hgnc_symbol). This is attempting using biomaRt's service and thus requires the optional biomaRt package to be installed.

Usage

.ensg_to_name(
  dat,
  ensg_col = "trait",
  new_col = "hgnc_symbol",
  build = "grch37"
)

Arguments

Value

Data.frame with appended column for names

Description

Prepare gwasvcf files for coloc. This method will extract SNPs from one file using one chrompos and then look up those SNPs in the other file – this is to ensure coloc can be conducted upon two datasets of different genomic builds without the need of liftover.

Usage

.gwasvcf_to_coloc_rsid(
  vcf1,
  vcf2,
  chrompos,
  type1 = NULL,
  type2 = NULL,
  build1 = "GRCh37",
  build2 = "GRCh37",
  verbose = TRUE
)

Arguments

Value

list of coloc-ready data, or NA if failed

Description

Write files for PWCoCo where data are read from two VCF objects or files.

Usage

.gwasvcf_to_pwcoco(vcf1, vcf2, chrompos, type1 = NULL, type2 = NULL, outfile)

Arguments

Value

0 if success, 1 if there was a problem

Description

Write files for PWCoCo where data are read from the OpenGWAS DB.

Usage

.ieugwasr_to_pwcoco(id1, id2, chrompos, type1 = NULL, type2 = NULL, outfile)

Arguments

Value

0 if success, 1 if there was a problem

Description

Calculates the inverse variance weighted delta method from the MendelianRandomization package

Usage

.ivw_delta(dat)

Arguments

Value

Results data.frame

Description

Helper function for message printing.

Usage

.print_msg(msg, verbose)

Arguments

Description

Sub-function to run PWCoCo

Usage

.pwcoco_sub(
  bfile,
  chrpos,
  pwcoco,
  maf = 0.01,
  p1 = 1e-04,
  p2 = 1e-04,
  p12 = 1e-05,
  workdir = tempdir(),
  verbose = TRUE
)

Arguments

Value

Results data.frame

Description

Helper function that is called from read_exposure and read_outcome. Extracts exposure and outcome data according to arguments. Should not be called directly.

Usage

.read_dataset(
  ids,
  rsids = NULL,
  pval = 5e-08,
  proxies = TRUE,
  proxy_rsq = 0.8,
  proxy_kb = 5000,
  proxy_nsnp = 5000,
  plink = NULL,
  bfile = NULL,
  clump_r2 = 0.01,
  clump_kb = 10000,
  pop = "EUR",
  type = "exposure",
  cores = 1,
  cores_proxy = 1,
  verbose = TRUE
)

Arguments

Value

Data.frame of datasets

See Also

[read_exposure()], [read_outcome()]

Description

Calculates the second term Taylor approximation for standard error of the Wald ratio method. From supplementary

Usage

.wr_taylor_approx(dat)

Arguments

Value

Results data.frame

Description

Annotates the data using given IDs

Usage

annotate_data(dat, id1, id2)

Arguments

Value

Data.frame of annotated dat

Description

Attempts to annotate disease names with EFO IDs using the 'epigraphdb' package. Note that the matching is fuzzy and some disease names will have multiple associated EFOs which may differ in definition slightly.

Usage

annotate_efo(dat, column = "outcome")

Arguments

Value

Data.frame with appended column for EFO IDs

Description

Attempts to convert ENSG IDs to gene names (hgnc_symbol). This is attempting using biomaRt's service and thus requires the optional biomaRt package to be installed.

Usage

annotate_ensg(
  dat,
  column = "exposure",
  gene_name_col = "hgnc_symbol",
  build = "grch37"
)

Arguments

Value

Data.frame with appended column for names

Description

Calculates portion of variance explained and F-statistic. If the data is lacking key information, i.e. allele frequencies, sample size or consists of only one SNP, then the approximate F-statistic will be used instead: $F = b ** 2 / SE ** 2$.

Usage

calc_f_stat(dat, f_cutoff = 10, force_approx = FALSE, verbose = TRUE)

Arguments

Value

Modified 'dat' data.frame (if f_cutoff > 0 supplied)

See Also

[do_mr()]

Description

Check if SNPs are good for use in analyses and mark them as such.

Usage

check_snps(dat, analyses = c("mr", "coloc"), drop = T)

Arguments

Details

List of analyses and what data are checked for:

• "MR"beta, SE, P value

• "coloc"chromosome, position, P value

Value

Data.frame

Description

Attempts to annotate SNPs as cis or trans depending on their location to the gene coding region. This is achieved using the 'biomaRt' R package.

Usage

cis_trans(
  dat,
  cis_region = 5e+05,
  chr_col = "chr.exposure",
  pos_col = "position.exposure",
  snp_col = NULL,
  values_col = "exposure",
  filter = "ensembl_gene_id",
  missing = "include",
  build = "grch37"
)

Arguments

Value

Data.frame with appended column for cis/trans status

Description

Combine MR and coloc results into one data.frame

Usage

combine_results(
  mr_res,
  coloc_res,
  mr_res.by = c("id.exposure", "id.outcome"),
  coloc_res.by = c("file.exposure", "file.outcome")
)

Arguments

Value

Data.frame of merged results

Description

Function to convert a data.frame to gwasvcf format.

Usage

dat_to_gwasvcf(
  dat,
  out,
  chr_col,
  pos_col,
  nea_col,
  ea_col,
  snp_col = NULL,
  eaf_col = NULL,
  beta_col = NULL,
  se_col = NULL,
  pval_col = NULL,
  n = NULL,
  n_case = NULL,
  name = NULL,
  bcf_tools = NULL,
  verbose = TRUE
)

Arguments

Value

gwasvcf object

See Also

[file_to_gwasvcf()] for converting files.

Description

Runs colocalisation using any of the following methods:

1. Coloc.abf, see coloc::coloc.abf()

2. Coloc.susie, see coloc::coloc.susie()

3. PWCoCo, see PWCoCo

NB: PWCoCo is not available on Windows.

Usage

do_coloc(
  dat,
  cdat = NA,
  method = "coloc.abf",
  coloc_window = 5e+05,
  plot_region = F,
  bfile = NULL,
  plink = NULL,
  pwcoco = NULL,
  workdir = tempdir(),
  cores = 1,
  verbose = TRUE
)

Arguments

Value

A data.frame of colocalistion results

See Also

[coloc::coloc.abf()], [coloc::coloc.susie()]

Description

Runs Mendelian randomisation and related analyses:

1. Wald ratio, see .wr_taylor_approx()

2. Inverse variance weighted, see .ivw.delta()

3. Steiger filtering, see TwoSampleMR::directionality_test()

Usage

do_mr(dat, f_cutoff = 10, all_wr = TRUE, verbose = TRUE)

Arguments

Value

A data.frame of MR results

See Also

[.wr_taylor_approx()], [.ivw_delta()], [TwoSampleMR::directionality_test()]

Description

Uses the Drug Genome Interaction DB's API to search for drug target-related evidence, including on: Druggable Genome, Clinically Actionable and Drug Resistant ontologies.

Usage

drug_target_evidence(dat, ensg_col = "exposure")

Arguments

Details

The lookup MUST be ENSG IDs.

Value

data.frame of results

Description

Extract SNPs based on region for colocalisation analyses. Can be used before calling the 'do_coloc' function or will be called as part of that function automatically.

Usage

extract_matched_regions(dat, window = 5e+05, cores = 1, verbose = TRUE)

Arguments

Value

Named list of matched regional data

See Also

[mrpipeline::do_coloc()]

Description

Function to convert a file (or files) to gwasvcf format.

Usage

file_to_gwasvcf(
  file,
  chr_col,
  pos_col,
  nea_col,
  ea_col,
  snp_col = NULL,
  eaf_col = NULL,
  beta_col = NULL,
  se_col = NULL,
  pval_col = NULL,
  n = NULL,
  n_case = NULL,
  name = NULL,
  header = TRUE,
  sep = "\t",
  cores = 1,
  bcf_tools = NULL,
  verbose = TRUE
)

Arguments

Value

gwasvcf object(s)

See Also

[dat_to_gwasvcf()] for converting data.frames

Description

Creates a forest plot of MR estimates from do_mr(). Will plot both the Wald ratios for all SNPs which form the instrument and inverse variance weighted method. However, if you wish for only the "discovery" results to be plotted (i.e. WR for single-SNP instruments and only IVW for multi-SNP instruments), then setting 'plot_all_res' to FALSE will achieve this. If the plot is too crowded, subsetting the results before passing them to this plotter will help.

Usage

forest_plot(
  res,
  snp_col = "snp",
  beta_col = "b",
  se_col = "se",
  pval_col = NULL,
  or_col = "or",
  or_lci_col = "or_lci95",
  or_uci_col = "or_uci95",
  method_col = "method",
  exposure_col = "exposure",
  outcome_col = "outcome",
  plot_all_res = TRUE
)

Arguments

Value

Plot

See Also

[do_mr()]

Description

Get column names from agnostic but formatted data.frame

Usage

get_col_name(df, data)

Arguments

Value

Name of column formatted for given data.frame

Description

Wrapper function for harmonise_data function in the 'TwoSampleMR' package.

Usage

harmonise(exposure, outcome, action = 1, cores = 1, verbose = TRUE)

Arguments

Value

Harmonised data.frame

See Also

[TwoSampleMR::harmonise_data()]

Description

Prints 'Hello, world!'.

Usage

hello()

Examples

hello()

Description

Check if SNP frequencies are ambiguous

Usage

is_ambiguous(freq, tol = 0.08)

Arguments

Value

True/False if ambiguous

Description

Check if SNP is palindromic

Usage

is_palindromic(A1, A2)

Arguments

Value

True/False if palindromic

Description

Entry point for the pipeline.

Usage

mr_pipeline(ids1, ids2, out_name = "", config_file = "", ...)

Arguments

Value

list of results for debugging

Description

In this function, an exposure-outcome pair are harmonised, analyses are ran on those data and those results are saved to a file. Analyses ran can be MR or colocalisation, as desired.

Usage

pairwise_analysis(
  exposure,
  outcome,
  res_path,
  ...,
  do_coloc = FALSE,
  cores = 1,
  verbose = TRUE
)

Arguments

Description

Read a dataset (or datasets) as exposure. Can accept both OpenGWAS IDs or file paths. Accepts only gwasvcf file formats. This function can clump data locally, if supplied with the 'plink' and 'bfile' arguments. If these are not supplied, clumping will take place on the OpenGWAS servers only for OpenGWAS IDs. If you would like to clump local files, please provide paths to Plink and bfiles.

Usage

read_exposure(
  ids,
  pval = 5e-08,
  plink = NULL,
  bfile = NULL,
  clump_r2 = 0.01,
  clump_kb = 10000,
  pop = "EUR",
  cores = 1,
  verbose = TRUE
)

Arguments

Value

Data.frame of exposure datasets

Description

Read a dataset (or datasets) as outcome. Can accept both OpenGWAS IDs or file paths. Accepts only gwasvcf file formats. This function can search for proxy SNPs locally, if supplied with the 'plink' and 'bfile' arguments. If these are not supplied, proxy searching will take place on the OpenGWAS servers only for OpenGWAS IDs. If you would like to search for proxies locally, please provide paths to Plink and bfiles.

Usage

read_outcome(
  ids,
  rsids,
  proxies = TRUE,
  proxy_rsq = 0.8,
  proxy_kb = 5000,
  proxy_nsnp = 5000,
  plink = NULL,
  bfile = NULL,
  cores = 1,
  cores_proxy = 1,
  verbose = TRUE
)

Arguments

Value

Data.frame of outcome datasets

Description

Plots a regional plot of the area being tested for colocalisation

Usage

regional_plot(
  dat,
  exposure,
  outcome,
  bfile = NULL,
  plink = NULL,
  verbose = TRUE
)

Arguments

Description

Validates parameters in config file

Usage

validate_config(conf)

Arguments

Value

Validated config class

Description

Creates a volcano plot of Wald ratios from [do_mr()]. This function will take all of the Wald ratios in the given data.frame and plot these. If the plot is too crowded, subsetting the results before passing them to this plotter will help. If 'plotly' is installed, the plot will be returned interactive.

Usage

volcano_plot(
  res,
  label = "outcome",
  snp_col = "snp",
  beta_col = "b",
  pval_col = "pval",
  or_col = "or",
  or_lci_col = "or_lci95",
  or_uci_col = "or_uci95",
  method_col = "method",
  force_static = FALSE
)

Arguments

Value

Plot

Description

If 'plotly' is installed, the plot will be returned interactive.

Usage

z_comparison_plot(dat1, dat2, z_col = "z", p_col = "pvalues", verbose = TRUE)

Arguments

Value

Plot

Description

Creates a Z-score plot for SNPs, where $Z = b / SE$. This should follow a parabollic shape and so can be used to find certain SNPs which may not follow this shape. If 'plotly' is installed, the plot will be returned interactive.

Usage

z_plot(
  dat,
  snp_col = "SNP",
  beta_col = "beta.exposure",
  se_col = "se.exposure",
  pval_col = "pval.exposure",
  force_static = FALSE
)

Arguments

Value

Plot