Per clone x window (1 Mb, or size): depth log2 of the pooled counts (log2((sum x + 0.5) / (sum e + 0.5))), relative to the clean reference pool's pseudobulk when ref_cells has >= min_ref cells, then
centred on the clone's autosomal median (states are relative to the clone baseline); SE = SD of the
per-cell window log2 ratios / sqrt(cells) (combined with the reference SE). Haplotype fraction: the
clone's cells are split into two random halves; each phase block is oriented on one half and counted on
the other, summed per window: nA, nB, n_hap, f, f_se (binomial SE x
the beta-binomial inflation 1 + (nbar - 1) * rho, nbar = molecule-weighted molecules per cell), and
n_blocks.
Arguments
- win
1 Mb window depth: list(windows (window, chrom, start, end), x, e) with x / e windows x cells
- votes, atoms
votes and atoms tables
- assign
data.table cell, clone (cells used for each clone's pseudobulk)
- ref_cells
reference pool cells (NULL = none)
- size
window size (bp; a multiple of 1 Mb), one value or one per clone (named)
- seed
RNG seed of the split halves
- cells
cell names in vote index order (default the columns of
win$x)- rho
beta-binomial overdispersion
- allelic_chroms
chromosomes with an allele channel (NULL = all)
- haploid
chromosomes with one copy in a normal cell (male chrX / chrY):
ploidy1- min_ref
minimal reference pool size