missing_by_sample {SNPfiltR}

Vizualise missing data per sample, remove samples above a missing data cutoff

Description

This function can be run in two ways: 1) Without 'cutoff' specified. This will vizualise the amount of missing data in each sample across a variety of potential missing data cutoffs. Additionally, it will show you a dotplot ordering the amount of overall missing data in each sample. Based on these visualizations, you can make an informed decision on what you think might be an optimal cutoff to remove samples that are missing too much data to be retained for downstream analyses. 2) with 'cutoff' specified. This option will show you the dotplot with the cutoff you set, and then remove samples above the missing data cutoff you set, and return the filtered vcf to you.

Usage

missing_by_sample(vcfR, popmap = NULL, cutoff = NULL)

Arguments

vcfR

a vcfR object

popmap

if specifies, it must be a two column dataframe with columns names 'id' and 'pop'. IDs must match the IDs in the vcfR object

cutoff

a numeric value between 0-1 specifying the maximum proportion of missing data allowed in a sample to be retained for downstream analyses

Details

Note: This decision is highly project specific, but these visualizations should help you get a feel for how very low data samples cannot be rescued simply by a missing data SNP filter. If you want to remove specific samples from your vcf that cannot be specified with a simple cutoff refer to this great tutorial which is the basis for the code underlying this function.

Value

if 'cutoff' is not specified, will return a dataframe containing the average depth and proportion missing data in each sample. If 'cutoff' is specified, the samples falling above the missing data cutoff will be removed, and the filtered vcfR object will be returned.

Examples

missing_by_sample(vcfR = SNPfiltR::vcfR.example)
missing_by_sample(vcfR = SNPfiltR::vcfR.example, cutoff = .7)