fit_logregr: Fit SINGLE's logistic regression

In rocioespci/single: Accurate consensus sequence from nanopore reads of a gene library

Fit SINGLE's logistic regression

Description

This is an auxiliary function in single package. It takes counts_pnq and for each position and nucleotide it fits SINGLE's logistic regression.

Usage

fit_logregr(
  counts_pnq,
  ref_seq,
  p_prior_errors,
  p_prior_mutations,
  save = FALSE,
  output_file_fits,
  output_file_data,
  verbose = FALSE,
  keep_fit_quality = FALSE
)

Arguments

counts_pnq	Data frame with columns position nucleoide quality counts, as returned by pileup_by_QUAL
ref_seq	DNAStringSet containing the true reference sequence.
p_prior_errors	Data frame with columns position nucleotide prior.error, as the one returned by p_prior_errors().
p_prior_mutations	Data frame with columns wt.base, nucleotide and p_mutation (probaility of mutation), as the one returned by p_prior_mutations().
save	Logical. Should data be saved in a output_file?
output_file_fits	File into which save the single fits if save=TRUE
output_file_data	File into which save the fitted data if save=TRUE
verbose	Logical.
keep_fit_quality	Logical. Should parameters related to the quality of the fit be returned in extra columns of the output?

Value

data.frame with columns position, nucleotide, slope and intercept (of the sigmoidal regression).

Examples

refseq_fasta <- system.file("extdata", "ref_seq.fasta", package = "single")
ref_seq = Biostrings::readDNAStringSet(refseq_fasta)
train_reads_example <- system.file("extdata", "train_seqs_500.sorted.bam",
                                   package = "single")
counts_pnq <- pileup_by_QUAL(bam_file=train_reads_example,
    pos_start=1,pos_end=10)
p_prior_mutations <- p_prior_mutations(rates.matrix = mutation_rate,
    mean.n.mut = 5,ref_seq = ref_seq)
p_prior_errors <- p_prior_errors(counts_pnq=counts_pnq)
fits <- fit_logregr(counts_pnq = counts_pnq,ref_seq=ref_seq,
    p_prior_errors = p_prior_errors,p_prior_mutations = p_prior_mutations)

rocioespci/single documentation built on April 18, 2023, 8:48 p.m.