oneByOnePrediction: Obtain a prediction for the ancestral clusters, removing...

In danjlawson/badMIXTURE: Validating Structure With Chromosome Painting

Description

Obtain the prediction of a mixture for each cluster (a P by P matrix), and return a matrix of this dimension but with replace the data for that cluster by its prediction, so that the diagonal is forced to match the prediction. This removes "genetic drift" specific to a given population.

Usage

oneByOnePrediction(mix, data.NbyP, coancestry.KbyP, pred.NbyPatK,
  tdend = NULL, clusterlist = NULL, ancestral = "solve")

Arguments

mix	A proposed mixture solution of dimension N by K; for example, as generated by STRUCTURE or ADMIXTURE.
data.NbyP	A matrix containing the data of dimension N by P, for example, as generated by ChromoPainter. the data must reflect similarity to P clusterings of the data.
coancestry.KbyP	A matrix of the cluster averaged data
pred.NbyPatK	A matrix of the cluster predictions
tdend	A dendrohram of the populations, used for ordering. Default: NULL, meaning use clusterlist
clusterlist	A list of populations, as described in compareMixtureToData. Default: NULL, meaning use tdend
ancestral	Ancestral model, as described in compareMixtureToData. Default: "solve"

Value

A matrix of dimension P by P

danjlawson/badMIXTURE documentation built on Sept. 27, 2019, 9:11 p.m.