Function that clusters genotypes into donors in an unsupervised manner (i.e. does not require the experimental design). This is useful if there is some doubt about the experimental design, or if one is not available.

Function that clusters genotypes into donors in an unsupervised manner (i.e. does not require the experimental design). This is useful if there is some doubt about the experimental design, or if one is not available.

construct_genotype_cluster_graph_unsupervised(
  file_locations,
  ncounts = 10,
  use_VAF = TRUE
)

Arguments

file_locations

the file locations

ncounts

numeric - the number of counts supporting each variant

use_VAF

if TRUE calculates genotype to genotype similarity on the basis of mean squared error between variant allele frequencies, if FALSE calculates relatedness (Pedersen et al. 2020) using a reimplementation of the Somalier approach.

Value

a list $graph_membership gives you cluster memberships $graph_plot gives a force directed embedding of the graph $membership_plot gives a heatmap of memberships $membership_matrix gives a matrix of channel memberships, $histogram gives a histogram of relatedness values used to cluster links into donors/generate a graph.

Examples

# \donttest{
construct_genotype_cluster_graph()
#> Error in construct_genotype_cluster_graph(): argument "file_locations" is missing, with no default
# }