QCreport: A post Alignment quality control of reads

In Roleren/ORFik: Open Reading Frames in Genomics

A post Alignment quality control of reads

Description

The ORFik QC uses the aligned files (usually bam files), fastp and STAR log files combined with annotation to create relevant statistics.

This report consists of several steps:
1. Convert bam file / Input files to ".ofst" format, if not already done. This format is around 400x faster to use in R than the bam format. Files are also outputted to R environment specified by envExp(df)
2. From this report you will get a summary csv table, with distribution of aligned reads and overlap counts over transcript regions like: leader, cds, trailer, lincRNAs, tRNAs, rRNAs, snoRNAs etc. It will be called STATS.csv. And can be imported with QCstats function.
3. It will also make correlation plots and meta coverage plots, so you get a good understanding of how good the quality of your NGS data production + aligner step were.
4. Count tables are produced, similar to HTseq count tables. Over mrna, leader, cds and trailer separately. This tables are stored as SummarizedExperiment, for easy loading into DEseq, conversion to normalized fpkm values, or collapsing replicates in an experiment. And can be imported with countTable function.

Everything will be outputed in the directory of your NGS data, inside the folder ./QC_STATS/, relative to data location in 'df'. You can specify new out location with out.dir if you want.
To make a ORFik experiment, see ?ORFik::experiment
To see some normal mrna coverage profiles of different RNA-seq protocols: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4310221/figure/F6/

Usage

QCreport(
  df,
  out.dir = resFolder(df),
  plot.ext = ".pdf",
  create.ofst = TRUE,
  complex.correlation.plots = TRUE,
  library.names = bamVarName(df),
  use_simplified_reads = TRUE,
  BPPARAM = bpparam()
)

Arguments

df	an ORFik experiment
out.dir	character, output directory, default: resFolder(df). Will make a folder within this called "QC_STATS" with all results in this directory. Warning: If you assign not default path, you will have a hazzle to load files later. Much easier to load count tables, statistics, ++ later with default. Update resFolder of df instead if needed.
plot.ext	character, default: ".pdf". Alternatives: ".png" or ".jpg". Note that in pdf format the complex correlation plots become very slow to load!
create.ofst	logical, default TRUE. Create ".ofst" files from the input libraries, ofst is much faster to load in R, for later use. Stored in ./ofst/ folder relative to experiment main folder.
complex.correlation.plots	logical, default TRUE. Add in addition to simple correlation plot two computationally heavy dots + correlation plots. Useful for deeper analysis, but takes longer time to run, especially on low-quality gpu computers. Set to FALSE to skip these.
library.names	character, default: bamVarName(df). Names to load libraries as to environment and names to display in plots.
use_simplified_reads	logical, default TRUE. For count tables and coverage plots a speed up for GAlignments is to use 5' ends only. This will lose some detail for splice sites, but is usually irrelevant. Note: If reads are precollapsed GRanges, set to FALSE to avoid recollapsing.
BPPARAM	how many cores/threads to use? default: bpparam(). To see number of threads used, do bpparam()$workers. You can also add a time remaining bar, for a more detailed pipeline.

Value

invisible(NULL) (objects are stored to disc)

See Also

Other QC report: QCplots(), QCstats()

Examples

# Load an experiment
df <- ORFik.template.experiment()
# Run QC
#QCreport(df, tempdir())
# QC on subset
#QCreport(df[9,], tempdir())

Roleren/ORFik documentation built on Nov. 13, 2024, 10 p.m.