For more details about the package or to cite, please visit https://www.biorxiv.org/content/10.1101/2023.05.07.539760v1.

MiSiPi.RNA

Characterization of small RNA pathways

Installation and Basic Usage

You can find the full documentation and examples here.

In order to install MiSiPi.RNA, you must first install devtools and BiocManager:

install.packages("devtools")

if (!require("BiocManager", quietly = TRUE))
  install.packages("BiocManager")
  
devtools::install_github("stupornova33/MiSiPi.RNA")

library(MiSiPi.RNA)

RNAfold

In order for this package to work, you must also have RNAfold and RNAplot (version >= 2.7.0) from the ViennaRNA package installed. See https://www.tbi.univie.ac.at/RNA/ for installation.

Optional dependencies

For converting the .ps/.eps output files from the siRNA/miRNA module to .png (see steps after "run_all" for examples), install ImageMagick and ghostscript.

Input

The input for MiSiPi.RNA's main function is an object created by the set_vars() function. Running set_vars will always be the first step in using this package. Below is a description of each of the parameters that will be passed to set_vars(). These should be changed based on your needs.

• roi - A bed file listing your regions of interest
• bam_file - A BAM file of aligned reads. Index file must also be present
• genome - A genome fasta file. Chromosome names must match the bed file
• plot_output - (TRUE or FALSE) If TRUE, MiSiPi.RNA will output plots as pdfs
• path_to_RNAfold - Full path to RNAfold executable
• path_to_RNAplot - Full path to RNAplot executable
• pi_pal - Palette option for the generated piRNA heatmap (see below)
• si_pal - Palette option for the generated siRNA heatmap (see below)
• annotate_region - (TRUE or FALSE) Plots annotated gene features below the hairpin arc plot which is useful for characterizing cisNAT loci
• weight_reads - Determines if read counts will be weighted. ("none", "locus_norm", "weight_by_prop", or an integer)
• gtf_file - Full path to a 9 column GTF file. Required only if annotate_region is TRUE. Note that the annotation column must contain features designated with "gene_id" or "transcript_id" (see https://www.ensembl.org/info/website/upload/gff.html?redirect=no#fields).
• write_fastas - (TRUE or FALSE) If TRUE, MiSiPi.RNA will write read pairs from functions to a file. Default is FALSE
• out_type - ("pdf" or "png") Specifies the output type. Default is "pdf"
• use_bed_names - (Optional) Specifies whether column 4 of the region of interest BED file should be used as names for output files (Default = FALSE)
• density_timeout - (Optional) The density plot and calc_phasing functions can take awhile to run for highly abundant loci, and thus we set a default timeout of 3600 seconds (1 hour). If you wish to override this, provide a time in seconds.

Palettes:

Palette options are:

• "RdYlBl"
• "BlYel"
• "yelOrRed"
• "MagYel"
• "Greens"

Example usage:

library(MiSiPi.RNA)

vars <- set_vars(
    roi = "path/to/bed",
    bam_file = "path/to/bam", 
    genome = "path/to/genome",
    plot_output = TRUE, 
    path_to_RNAfold = "path/to/ViennaRNA/RNAfold.exe",
    path_to_RNAplot = "path/to/ViennaRNA/RNAplot.exe",
    pi_pal = "BlYel",
    si_pal = "RdYlBl",
    annotate_region = TRUE,
    weight_reads = "none",
    gtf_file = "path/to/gtf",
    write_fastas = FALSE,
    out_type = "pdf", 
    use_bed_names = FALSE,
    density_timeout = 3600
)



# run only for siRNAs
misipi_rna(vars, method = "siRNA")

The method parameter determines if your files will be processed for MicroRNA ("miRNA"), Piwi-interacting RNA ("piRNA"), Small interferring RNA ("siRNA"). By default, misipi_rna runs all three methods and combines the output plots.

Running with "all" method:

# run all methods
misipi_rna(vars)

In addition to processing files for miRNA, piRNA, and siRNA, the "all" method outputs a table with metrics and statistics which can be used for summarization or machine learning. See the documentation for more details regarding values in table.

Example Data Provided with MiSiPi.RNA

See examples/examples.Rmd for more.

library(MiSiPi.RNA)

roi <- system.file("extdata", "dmel_mir-bantam.bed", package = "MiSiPi.RNA")
bam_file <- system.file("extdata", "dmel_bantam.bam", package = "MiSiPi.RNA")
genome <- system.file("extdata", "dmel_chr3L.fasta", package = "MiSiPi.RNA")
annot <- system.file("extdata", "dmel_example.gtf", package = "MiSiPi.RNA")
path_to_RNAfold = "" # Replace with your path to RNAfold and RNAplot here
path_to_RNAplot = ""

vars <- set_vars(
  roi = roi,
  bam_file = bam_file,
  genome = genome,
  path_to_RNAfold = path_to_RNAfold,
  path_to_RNAplot = path_to_RNAplot,  
  pi_pal = "BlYel",
  si_pal = "RdYlBl",
  annotate_region = TRUE,
  gtf_file = annot,
  out_type = "png")

misipi_rna(vars, method = "miRNA")

To convert .ps/.eps files to .png

eps2png(path_to_magick_exe, file_dir)

where path_to_magick_exe is the full path to the binary executable and file_dir is the folder containing the files to convert. This will also be the output folder.

Use built-in machine learning model to characterize loci:

See full documentation for more details.

ml_probability(path_to_tables = "full/path/to/run_all/", table = "table_ml.txt")

The path_to_tables parameter is the path to the directory that was created when misipi_rna() was run with the default method. It will be called "run_all/".

To make a plot summary and sortable table of results:

This function generates one html file based on the chosen class type. Use "type =" to choose one of the following classes: miRNA, piRNA, or siRNA.

make_html_summary(path_to_tables = "full/path/to/run_all/", type = "miRNA", ml_plots = FALSE)

The path_to_tables parameter is the path to the directory that was created when misipi_rna() was run with the default method. It will be called "run_all/".

The ml_plots parameter is intended for users that have already run ml_probability(). The default for this option is false.