clean_fq: Removes the noise of an individual fastq file

In thierrygosselin/stackr: Run stacks software pipeline for RADseq analysis inside R

Description

This function reads the fastq file of an individual and clean it by removing:

• unique reads with high coverage (likely paralogs or TE)

• distinct reads with low coverage

Usage

clean_fq(
  fq.files,
  min.coverage.threshold = 2L,
  max.coverage.threshold = "high.coverage.unique.reads",
  remove.unique.reads = TRUE,
  write.blacklist = TRUE,
  write.blacklist.fasta = TRUE,
  compress = FALSE,
  output.dir = NULL,
  parallel.core = parallel::detectCores() - 1
)

Arguments

fq.files	(character, path). The path to the individual fastq file to check. Default: fq.files = "my-sample.fq.gz".
min.coverage.threshold	(integer). Minimum coverage threshold. The function will remove distinct reads with coverage <= to the threshold. To turn off, min.coverage.threshold = NULL or 0L. Default: min.coverage.threshold = 2L.
max.coverage.threshold	(integer, character). Maximum coverage threshold. The function will remove distinct reads with coverage >= than this threshold. To turn off, max.coverage.threshold = NULL. The default, use the starting depth where high coverage unique reads are observed. Default: max.coverage.threshold = "high.coverage.unique.reads".
remove.unique.reads	(logical). Remove distinct unique reads with high coverage. Likely paralogs or Transposable elements. Default: remove.unique.reads = TRUE.
write.blacklist	(logical). Write the blacklisted reads to a file. Default: write.blacklist = FALSE.
write.blacklist.fasta	(logical). Write the blacklisted reads to a fasta file. Default: write.blacklist.fasta = FALSE.
compress	(logical) To compress the output files. If you have the disk space, don't compress, it's way faster this way to write. Default: compress = FALSE.
output.dir	(path) Write the cleaned fq files in a specific directory. Default: output.dir = NULL, uses the working directory.
parallel.core	(integer) Enable parallel execution with the number of threads. Default: parallel.core = parallel::detectCores() - 1.

Details

coming soon, just try it in the meantime...

Value

The function returns a cleaned fq file with the name of the sample and -cleaned appended to the filename.

Examples

## Not run: 
require(vroom)

# for one sample
clean.id <- stackr::clean_fq(
  fq.files = "my-sample.fq.gz",
  min.coverage.threshold = 7L,
  max.coverage.threshold = "high.coverage.unique.reads"
  )

# for multiple samples in parallel
# require(progressr)

 progressr::with_progress({
   clean <- stackr::clean_fq(
     fq.files = 04_process_radtags,
      min.coverage.threshold = 2L,
      max.coverage.threshold = "high.coverage.unique.reads",
      write.blacklist = TRUE,
      write.blacklist.fasta = TRUE,
      compress = FALSE,
      output.dir = "04_process_radtags/cleaned_fq"
 )
 })

## End(Not run)

thierrygosselin/stackr documentation built on Nov. 11, 2020, 11 a.m.