Single chromosome run imputation of small, NA read gaps in fragCounter coverage

Run imputation of small, NA read gaps in fragCounter using partition regression of signal in neighborhood adjacent to the gap. The gaps are typically small, less than 500 bp on average and inter spaced among large segments of continuous, homogeneous signal. The gap signal is far less than expected for real deletions, even homozygous events Imputation greatly improves the foreground signal in downstream tools like JaBbA and GISTIC.

Note, this function was designed to be run as part of the get_cov_gap_imputation() workflow. Also, see chrompar() for executing in parallel.

Usage

get_imputed_gaps_per_chromosome(
  chrom_for_imputation,
  gapless_coverage_whitelist,
  coverage_gaps,
  sample_id,
  make_plots = F,
  path_for_plots = NULL
)

Arguments

chrom_for_imputation

The chromosome to run gap imputation algorithm on

gapless_coverage_whitelist

GenomicRanges object of coverage after removing blacklist regions and gaps

coverage_gaps

GenomicRanges object of gaps from the original coverage

sample_id

Name of sample

make_plots

Generate diagnostic plots of imputation results? default: FALSE

path_for_plots

Path to directory to hold diagnostic plot small PNGs (~80 KB per plot)

Value

data.table with gap imputed coverage