rank_clones: rank clones

In SingerLab/gac: Genetic Analysis of Cells

rank clones

Description

STILL IN DEVELOPMENT: UNTESTED

Usage

rank_clones(cnr, cluster_column, rank_by, fga.method, prefix = NULL)

Arguments

cnr	a cnr object
cluster_column	which cluster column to use; must be one of "BrayC", "ConsensusC", or "final_cluster"
rank_by	option to rank, 'fga' to rank by genomic complexity based on the fraction of genome altered. Or 'frequency', ranked from most to least least prevalent.
fga.method	option to estimate cluster FGA, either 'genomic' or 'proxy'. In method 'genomic' the altered bins are weighed by their genomic length in base pairs (bp), and the fraction is estimated based on the total genome length. Method 'proxy' does not weigh by genomic lenght, instead it approximates FGA by estimating the proportion of altered bins. Alterations are deviations from a diploid state e.g. != 2.
prefix	prefix charcter to append to Consensus Clusters

Value

rank_clones returns a ranked_cluster vector to 'Y'. The ranking of the clone is user specified by frequency or genomic complexity. When ranking clones by frequency, the specified clusters are ranked in descending order, where the most frequent clone is assigned clone '1', and all less frequent clones are assigned a numerical label in sequential order. Ties are broke at random. When ranking by genomic complexity, then the fraction of the genome altered (FGA) is computed based on a diploid genome (2n). And ranked in ascending order, where the least altered cell is clone '1' and more altered clones are assigned a numerical label in sequential order. Ties are broken at random.

Clusters with 3 or more cells are sumarized using the same parameters as in get_cluster_profiles. However, for clusters with 2 cells, are sumarized by taking the 'max' at each bin, and those with one cell, then the cell was used.

SingerLab/gac documentation built on March 23, 2024, 5:15 a.m.