calc_cluster_order_probs: Calculate for a pair of clusters whether a has a higher CCF...

In Wedge-Oxford/dpclust: DPClust - An R package for subclonal reconstruction of tumours using sequencing data

Calculate for a pair of clusters whether a has a higher CCF than b.

Description

This function calculates cluster order probabilities by obtaining mutation preferences throughout the MCMC iterations for each provided cluster and then classifies a pair of clusters in groups: Greater than / equal (GT-EQ), less than / equal (LT-EQ), greater than (GT), less than (LT), equal (EQ) or undertain (uncertain). These classifications are obtained by sampling pairs of SNVs from either cluster and account how often SNV 1 is assigned a higher CCF in the preferences than SNV 2.

Usage

calc_cluster_order_probs(GS.data, density, mut_assignments, clusterids,
  cluster_ccfs, no.muts, no.timepoints, no.iters, no.iters.burn.in,
  no.samples = 1000)

Arguments

GS.data	MCMC output with assignments and cluster locations
clusterids	Clusterids to run through
no.muts	The total number of mutations
no.timepoints	Total number of samples in this dataset
no.iters	Total number of iterations
no.iters.burn.in	Number of iterations to use as burn-in

Value

A array multi-dimensional array with in each cell whether the column cluster has a higher CCF than the row cluster across the samples in the third dimension

Author(s)

sd11 Note: This approach only works with the density based mutation assignment strategy Should return classification of each cluster/cluster pair (row vs column): * GT / LT / EQ / GT-EQ / EQ-LT / uncertain / NA

Wedge-Oxford/dpclust documentation built on July 6, 2024, 2:02 p.m.