summarise_window_parallelism: Summarise Parallel Evolution in Windows with Significant...

In JimWhiting91/afvaper: AF-vapeR: Allele Frequency Vector Analysis of Parallel Evolutionary Responses

Summarise Parallel Evolution in Windows with Significant Eigenvalues

Description

Produces a summary table describing how replicate population pairs load on relevant eigenvectors. This information allows you to make sense of what patterns are driving large eigenvalues. For each significant window, the table summarises how many replicates load (above a user-defined threshold) onto the relevant eigenvector, and whether they load in the same direction (parallel) or opposite direction (anti-parallel). For significant windows on eigenvectors 2+, the table includes information on all higher order eigenvectors (e.g. for eigenvector 2, there is a row for both eigenvector 1 and 2). This is because the analysis is focussed on the sum of eigenvalues, therefore all higher order eigenvalues are relevant to interpretation.

Usage

summarise_window_parallelism(
  window_id,
  eigen_res,
  loading_cutoff = 0.3,
  eigenvector = 1
)

Arguments

window_id	A character string corresponding to a single focal chromosomal window, or a vector of multiple windows. These windows must be in the output of names(eigen_res). Significant windows can be found using the function signif_eigen_windows().
eigen_res	A list of results from eigen_analyse_vectors(). This can be easily created by combining eigen_analyse_vectors with lapply. Each element of the input list should correspond to the eigen decomposition results for a given genome window.
loading_cutoff	Numeric value corresponding to the loading threshold that replicate population pairs need to exceed to be considered as associated with a given eigenvector. This threshold applies to positive or negative loadings as absolute loadings are considered when comparing to the threshold. The threshold can be considered as the correlation coefficient between the eigenvector scores for a given replicate population pair, and its observed allele frequency change values. As such, higher loading thresholds correspond to a stronger agreement, and closer association, between a replicate population pair and the relevent eigenvector.
eigenvector	Integer value for the focal eigenvector to be considered, e.g. 1 = eigenvector 1

Value

A data.frame table summarising parallel evolution

JimWhiting91/afvaper documentation built on Aug. 30, 2022, 10:03 a.m.