Filter markers based on genotyping / missing rate

Filter markers based on their genotyping (call) rate, i.e. the proportion of non-missing genotypes per marker. This is a convenient way to remove SNPs with too much missing data before downstream analyses.

Filter targets: SNPs.

Statistics: marker-level missingness (MISSING_PROP from generate_stats).

filter_genotyping(
  data,
  interactive.filter = TRUE,
  filter.genotyping = NULL,
  filename = NULL,
  parallel.core = parallel::detectCores() - 1,
  verbose = TRUE,
  ...
)

Arguments

data

A Genomic Data Structure (GDS) file or connection (SeqVarGDSClass or gds.file) created by read_vcf or other radiator import functions.

interactive.filter

(optional, logical). Do you want the filtering session to be interactive ? With interactive.filter = TRUE, helper tables and plots are shown, and the user is prompted for a threshold. Default: interactive.filter = TRUE.

filter.genotyping

(optional, character or double). Two modes are available:

• character string: filter.genotyping = "outliers" uses the upper outlier value from the missingness boxplot as threshold (i.e. removes high-missingness outliers);

• double: filter.genotyping = 0.2 allows up to 0.2 missing genotypes (i.e. keeps markers with MISSING_PROP <= 0.2).

Default: filter.genotyping = NULL.

If interactive.filter = FALSE and filter.genotyping = NULL, the function returns data unchanged.

filename

(optional, character). Basename for some output files (plots, tables). If NULL, a name is generated automatically from the function and date. Default: filename = NULL.

parallel.core

(optional, integer). Number of CPU cores to use in helper statistics. Default: parallel.core = parallel::detectCores() - 1.

verbose

(optional, logical). Show messages and progress. Default: verbose = TRUE.

...

(optional) Advance mode that allows to pass further arguments for fine-tuning the function. Also used for legacy arguments (see details or special section)

Value

The filtered dataset, of the same type as the input:

• if data is a GDS connection (SeqVarGDSClass), the same connection is returned with updated markers.meta;

• if data is a GDS file path, the path is returned (the file is updated on disk).

Side-effects:

• helper tables and plots are written under filter_genotyping_YYYYMMDD@HHMM;

• markers.meta inside the GDS is updated and synchronised;

• the filters.parameters file is updated to record the filter.

Details

Internally, the function uses generate_stats to compute marker-level missingness (MISSING_PROP). Markers with missing proportion greater than the chosen threshold are blacklisted and have their FILTERS column updated to "filter.genotyping" in markers.meta.

With interactive.filter = TRUE, a helper table and plot are generated to assist in choosing a threshold:

• genotyping.helper.table.tsv: number of markers kept/removed for thresholds from 0 to 1 by 0.1;

• if strata are present, markers.pop.missing.helper.table.tsv summarises missingness per population;

• a PDF/PNG helper plot summarising these patterns.

This function operates on the GDS representation of the data and is meant for genotyping/missingness-based pruning after VCF import (e.g. after read_vcf), in contrast to VCF-level slimming using tools like filter_monomorphic_vcf.

See also

filter_common_markers, filter_monomorphic, filter_ma, read_vcf, filter_rad.

Author

Thierry Gosselin thierrygosselin@icloud.com

Examples

if (FALSE) { # \dontrun{
# Filter markers with more than 20% missing genotypes
gds <- radiator::read_vcf(data = "populations.snps.vcf") %>%
       radiator::filter_genotyping(
            data               = .,
            interactive.filter = FALSE,
            filter.genotyping  = 0.2
)
} # }