Skip to contents

Per chromosome: tabix <file> <chrom> | cut -f2,3,4,7,11,13 read with fread(cmd = ) (an awk filter keeps hap 1 / 2 with ps > 0 before R sees the rows), or read_fragments() when the tabix binary or index is missing; fragments of cells with haplotype A / B and a phase set are collapsed into molecules (collapse_molecules()), each molecule is assigned to the atom containing its start and the votes are aggregated per cell x atom x phase set. Memory stays per chromosome (integer cell indices, gc() between chromosomes). A failing tabix command stops with its error; a chromosome with atoms but no haplotype-assigned fragment of cells gives a warning.

Usage

read_votes(
  frag_file,
  cells,
  atoms,
  chroms = NULL,
  gap = 2000,
  tabix = caller_defaults()$tabix,
  split_ps = FALSE,
  clean = FALSE
)

Arguments

frag_file

fragments.tsv.gz (+ .tbi)

cells

cell barcodes; the votes' cell is the index into this vector

atoms

atoms_from_regions() output

chroms

chromosomes to read (default: the atoms' chromosomes)

gap

molecule gap (bp)

tabix

path of the tabix binary ("" = the tabix on the PATH; NULL / not found = the R fallback through Rsamtools)

split_ps

split molecules by phase set (caller parameter mol_split_ps; default FALSE)

clean

votes from clean (not blacklisted) fragments only (caller parameter votes_clean; default FALSE)

Value

votes data.table cell, atom, ps, nA, nB (passes check_votes())

Details

clean = TRUE keeps only the fragments the extractor counts in the depth layers (no blacklist bit in flags, the extractor's clean = (flags & F_BLACKLIST) == 0), in the tabix path (awk) and in the fallback; split_ps = TRUE gives one vote per molecule x phase set (collapse_molecules()).