AntibodyForests (version 1.1.0)

Af_compare_methods: Function to compare trees created with different algorithms from the same clonotype.

Description

Function to compare different trees from the same clonotype to compare various graph construction and phylogenetic reconstruction methods.

Usage

Af_compare_methods(
  input,
  min.nodes,
  include.average,
  distance.method,
  depth,
  clustering.method,
  visualization.methods,
  parallel,
  num.cores
)

Value

A list with all clonotypes that pass the min.nodes threshold including the distance matrix, possible clustering and visualization

Arguments

input

A list of AntibodyForests-objects as output from the function Af_build(). These objects should contain the same samples/clonotypes. For easy interpretation of the results, please name the objects in the list according to their tree-construction method.

min.nodes

The minimum number of nodes in a tree to include in the comparison, this includes the germline. Default is 2 (this includes all trees).

include.average

If TRUE, the average distance matrix and visualizations between the trees is included in the output (default FALSE)

distance.method

The method to calculate the distance between trees (default euclidean) 'euclidean' : Euclidean distance between the depth of each node in the tree 'GBLD' : Generalized Branch Length Distance, derived from Mahsa Farnia & Nadia Tahiri, Algorithms Mol Biol 19, 22 (2024). https://doi.org/10.1186/s13015-024-00267-1

depth

If distance.methods is 'euclidean', method to calculate the germline-to-node depth (default edge.count) 'edge.count' : The number of edges between each node and the germline 'edge.length' : The sum of edge lengths between each node and the germline

clustering.method

Method to cluster trees (default NULL) NULL : No clustering 'mediods' : Clustering based on the k-mediods method. The number of clusters is estimated based on the optimum average silhouette.

visualization.methods

The methods to analyze similarity (default NULL) NULL : No visualization 'PCA' : Scatterplot of the first two principal components. 'MDS' : Scatterplot of the first two dimensions using multidimensional scaling. "heatmap' : Heatmap of the distance

parallel

If TRUE, the depth calculations are parallelized across clonotypes (default FALSE)

num.cores

Number of cores to be used when parallel = TRUE. (Defaults to all available cores - 1)

Examples

plot <- Af_compare_methods(input = list("Default" = AntibodyForests::af_default,
                                        "MST" = AntibodyForests::af_mst,
                                        "NJ" = AntibodyForests::af_nj),
                           depth = "edge.count",
                           visualization.methods = "heatmap",
                           include.average = TRUE)
plot$average