Evaluate results from PLINK PCA on combined study and reference data

Source: R/individualQC.R

Evaluates and depicts results of plink --pca (via run_check_ancestry or externally conducted pca) on merged genotypes from individuals to be QCed and individuals of reference population of known genotypes. Currently, check ancestry only supports automatic selection of individuals of European descent. It uses information from principal components 1 and 2 returned by plink --pca to find the center of the European reference samples (mean(PC1_europeanRef), mean(PC2_europeanRef). It computes the maximum Euclidean distance (maxDist) of the European reference samples from this centre. All study samples whose Euclidean distance from the centre falls outside the circle described by the radius r=europeanTh* maxDist are considered non-European and their IDs are returned as failing the ancestry check. check_ancestry creates a scatter plot of PC1 versus PC2 colour-coded for samples of the reference populations and the study population.

evaluate_check_ancestry(
  indir,
  name,
  prefixMergedDataset,
  qcdir = indir,
  europeanTh = 1.5,
  defaultRefSamples = c("HapMap", "1000Genomes"),
  refSamples = NULL,
  refColors = NULL,
  refSamplesFile = NULL,
  refColorsFile = NULL,
  refSamplesIID = "IID",
  refSamplesPop = "Pop",
  refColorsColor = "Color",
  refColorsPop = "Pop",
  studyColor = "#2c7bb6",
  refPopulation = c("CEU", "TSI"),
  legend_labels_per_row = 6,
  legend_text_size = 5,
  legend_title_size = 7,
  axis_text_size = 5,
  axis_title_size = 7,
  title_size = 9,
  highlight_samples = NULL,
  highlight_type = c("text", "label", "color", "shape"),
  highlight_text_size = 3,
  highlight_color = "#c51b8a",
  highlight_shape = 17,
  highlight_legend = FALSE,
  interactive = FALSE,
  verbose = FALSE
)

Arguments

indir

[character] /path/to/directory containing the basic PLINK data files name.bim, name.bed, name.fam files.
name

[character] Prefix of PLINK files, i.e. name.bed, name.bim, name.fam.
prefixMergedDataset

[character] Prefix of merged dataset (study and reference samples) used in plink --pca, resulting in prefixMergedDataset.eigenvec.
qcdir

[character] /path/to/directory/with/QC/results containing prefixMergedDataset.eigenvec results as returned by plink --pca. Per default qcdir=indir.
europeanTh

[double] Scaling factor of radius to be drawn around center of European reference samples, with study samples inside this radius considered to be of European descent and samples outside this radius of non-European descent. The radius is computed as the maximum Euclidean distance of European reference samples to the centre of European reference samples.
defaultRefSamples

[character] Option to use pre-downloaded individual and population identifiers from either the 1000Genomes or HapMap project. If refSamples and refSamplesFile are not provided, the HapMap identifiers (or 1000Genomes is specified) will be used as default and the function will fail if the reference samples in the prefixMergedDataset do not match these reference samples. If refColors and refColorsFile are not provided, this also sets default colors for the reference populations.
refSamples

[data.frame] Dataframe with sample identifiers [refSamplesIID] corresponding to IIDs in prefixMergedDataset.eigenvec and population identifier [refSamplesPop] corresponding to population IDs [refColorsPop] in refColorsfile/refColors. If refSamples and refSamplesFile are not specified, defaultRefSamples will be used as reference.
refColors

[data.frame, optional] Dataframe with population IDs in column [refColorsPop] and corresponding colour-code for PCA plot in column [refColorsColor]. If refColors and refColorsFile are not specified and refSamples and refSamplesFile are not specified, default colors will be determined from the defaultRefSamples option. If refColors and refColorsFile are not specified and but refSamples or refSamplesFile are given, ggplot default colors will be used.
refSamplesFile

[character] /path/to/File/with/reference samples. Needs columns with sample identifiers [refSamplesIID] corresponding to IIDs in prefixMergedDataset.eigenvec and population identifier [refSamplesPop] corresponding to population IDs [refColorsPop] in refColorsfile/refColors. If both refSamplesFile and refSamples are not NULL, defaultRefSamples information is used.
refColorsFile

[character, optional] /path/to/File/with/Population/Colors containing population IDs in column [refColorsPop] and corresponding colour-code for PCA plot in column [refColorsColor]. If refColors and refColorsFile are not specified and refSamples and refSamplesFile are not specified, default colors will be determined from the defaultRefSamples option. If refColors and refColorsFile are not specified and but refSamples or refSamplesFile are given, ggplot default colors will be used.
refSamplesIID

[character] Column name of reference sample IDs in refSamples/refSamplesFile.
refSamplesPop

[character] Column name of reference sample population IDs in refSamples/refSamplesFile.
refColorsColor

[character] Column name of population colors in refColors/refColorsFile
refColorsPop

[character] Column name of reference sample population IDs in refColors/refColorsFile.
studyColor

[character] Colour to be used for study population if plot is TRUE.
refPopulation

[vector] Vector with population identifiers of European reference population. Identifiers have to be corresponding to population IDs [refColorsPop] in refColorsfile/refColors.
legend_labels_per_row

[integer] Number of population names per row in PCA plot.
legend_text_size

[integer] Size for legend text.
legend_title_size

[integer] Size for legend title.
axis_text_size

[integer] Size for axis text.
axis_title_size

[integer] Size for axis title.
title_size

[integer] Size for plot title.
highlight_samples

[character vector] Vector of sample IIDs to highlight in the plot (p_ancestry); all highlight_samples IIDs have to be present in the IIDs of the prefixMergedDataset.fam file.
highlight_type

[character] Type of sample highlight, labeling by IID ("text"/"label") and/or highlighting data points in different "color" and/or "shape". "text" and "label" use ggrepel for minimal overlap of text labels ("text) or label boxes ("label"). Only one of "text" and "label" can be specified.Text/Label size can be specified with highlight_text_size, highlight color with highlight_color, or highlight shape with highlight_shape.
highlight_text_size

[integer] Text/Label size for samples specified to be highlighted (highlight_samples) by "text" or "label" (highlight_type).
highlight_color

[character] Color for samples specified to be highlighted (highlight_samples) by "color" (highlight_type).
highlight_shape

[integer] Shape for samples specified to be highlighted (highlight_samples) by "shape" (highlight_type). Possible shapes and their encoding can be found at: https://ggplot2.tidyverse.org/articles/ggplot2-specs.html#sec:shape-spec
highlight_legend

[logical] Should a separate legend for the highlighted samples be provided; only relevant for highlight_type == "color" or highlight_type == "shape".
interactive

[logical] Should plots be shown interactively? When choosing this option, make sure you have X-forwarding/graphical interface available for interactive plotting. Alternatively, set interactive=FALSE and save the returned plot object (p_ancestry) via ggplot2::ggsave(p=p_ancestry, other_arguments) or pdf(outfile) print(p_ancestry) dev.off().
verbose

[logical] If TRUE, progress info is printed to standard out.

Value

Named [list] with i) fail_ancestry, containing a [data.frame] with FID and IID of non-European individuals and ii) p_ancestry, a ggplot2-object 'containing' a scatter plot of PC1 versus PC2 colour-coded for samples of the reference populations and the study population, which can be shown by print(p_ancestry) and iii) plot_data, a data.frame with the data visualised in p_ancestry (ii).

Details

Both run_check_ancestry and evaluate_check_ancestry can simply be invoked by check_ancestry.

1000 Genomes samples were downloaded from https://www.internationalgenome.org/category/sample/, HapMap Phase 3 samples were downloaded from https://www.broadinstitute.org/medical-and-population-genetics/hapmap-3.

Examples

if (FALSE) {
qcdir <- system.file("extdata", package="plinkQC")
name <- "data"

# whole dataset
fail_ancestry <- evaluate_check_ancestry(indir=qcdir, name=name,
refSamplesFile=paste(qcdir, "/HapMap_ID2Pop.txt",sep=""),
refColorsFile=paste(qcdir, "/HapMap_PopColors.txt", sep=""),
prefixMergedDataset="data.HapMapIII", interactive=FALSE)

# highlight samples
highlight_samples <- read.table(system.file("extdata", "keep_individuals",
package="plinkQC"))
fail_ancestry <- evaluate_check_ancestry(indir=qcdir, name=name,
refSamplesFile=paste(qcdir, "/HapMap_ID2Pop.txt",sep=""),
refColorsFile=paste(qcdir, "/HapMap_PopColors.txt", sep=""),
prefixMergedDataset="data.HapMapIII", interactive=FALSE,
highlight_samples = highlight_samples[,2],
highlight_type = c("text", "shape"))
}