evaluate_bias: Evaluate positional contributions to RPF counts

In amandamok/choros: Bias Correction for Ribosome Profiling Data

Evaluate positional contributions to RPF counts

Description

This function will perform a series of leave-one-out regressions to evaluate the contribution of individual positions to overall RPF count.

Usage

evaluate_bias(
  dat,
  which_column = "count",
  transcripts_fa_fname,
  transcripts_length_fname,
  trunc5 = 20,
  trunc3 = 20,
  num_f5_codons = 6,
  num_f3_codons = 6,
  type = "codon",
  metric = "corr"
)

Arguments

dat	data frame containing regression predictors and column of RPF counts
which_column	character; name of column containing counts
transcripts_fa_fname	character; file path to transcriptome .fasta file
transcripts_length_fname	character; file path to transcriptome lengths file
trunc5	integer; number of codons (starting at start codon) to omit in codon correlation regression
trunc3	integer; number of codons (starting at stop codon) to omit in codon correlation regression
num_f5_codons	integer; number of codons 5' of A site to include in regression
num_f3_codons	integer; number of codons 3' of A site to include in regression
type	character; one of 'codon' or 'nt'
metric	character; one of 'corr' or 'norm'

Value

A named numeric vector of positional contributions to RPF count. Values will correspond to either the decrease in correlation between predicted and true RPF counts in the leave-one-out model (as computed in the 'iXnos' model), or the norm of regression coefficients at that position.

amandamok/choros documentation built on March 15, 2023, 7:57 p.m.