README.md

In amandamok/choros: Bias Correction for Ribosome Profiling Data

choros

choros is a software pipeline to estimate and correct the sequence biases present in ribosome profiling datasets. A preprint describing the method in greater detail can be found on bioRxiv.

Briefly, choros performs the following:

1. Annotation of RPF alignments with: codon index corresponding to the ribosome A-site codon position; codons present in the ribosome A-, P-, and E-site positions; RPF 5' and 3' digest lengths (sequence length between RPF ends and A site); and RPF 5' and 3' bias sequences (nucleotides at RPF ends).

2. Generation of a training dataset of RPF counts on transcripts with high RPF density for regression modeling.

3. Fitting of a negative binomial regression, modeling RPF counts on A-/P-/E-site codon identities, RPF 5' and 3' digest lengths, 5' and 3' bias sequences, and RPF GC-content.

4. Bias correction using scaling factors computed from regression modeling.

5. Computation and visualization of feature importance measurements per codon or nucleotide position, as performed in iXnos to evaluate sequence biases before and after bias correction.

Installation

1. Install R

2. Install dependencies

Packages from CRAN:

install.packages(c("parallel", "doParallel", "foreach", "reshape", 
                   "ggplot2", "MASS", "patchwork", "RColorBrewer", 
                   "data.table", "devtools"))

Packages from Bioconductor:

if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("Rsamtools")
BiocManager::install("Biostrings")

1. Install choros

devtools::install_github("lareaulab/choros", ref = "main",
                         build_vignettes = TRUE)

Input files

• A .bam file of alignments of pre-processed ribosome protected fragments (RPFs)
• A FASTA file of transcript sequences (including 5' and 3' UTR regions)
• A transcriptome lengths file containing 5' UTR, CDS, and 3' UTR lengths
• An A-site offset rules file containing A-site offsets per read length and frame

Running choros

A vignette showing usage and output of choros on a small sample dataset is included:

vignette("choros_vignette")

Pre-processing

choros assumes that all read pre-processing steps have been performed by the user prior to regression modeling. These steps may include:

• adapter trimming
• filtering out of contaminant reads (i.e. ribosomal and noncoding RNAs)
• alignment to the transcriptome

In the case of reads with UMIs, the authors recommend additional UMI deduplication steps. See the choros paper github repo for example scripts.

The authors additionally recommend computing posterior mapping weights for multi-mapping reads with RSEM.

Installation notes

| | cluster | macbook pro | | :--- | :--- | :--- | | time to install | 1 min (if other dependencies are installed) | < 1 min (skipping other dependencies) | | time to run test data | 3 min | 3 min | | system version | ubuntu 18.04.5 (linux 4.15.0-141-generic) | macOS 12.3.1 | | R version | 3.4.4 | 4.1.2 | | | | | | parallel | 3.4.4 | 4.1.2 | | doParallel | 1.0.17 | 1.0.17 | | foreach | 1.5.2 | 1.5.2 | | reshape | 0.8.9 | 0.8.9 | | Rsamtools | 1.30.0 | 2.10.0 | | ggplot2 | 3.3.6 | 3.4.1 | | MASS | 7.3-54 | 7.3-58.2 | | Biostrings | 2.46.0 | 2.62.0 | | patchwork | 1.1.1 | 1.1.2 | | RColorBrewer | 1.1-3 | 1.1-3 | | data.table | 1.14.2 | 1.14.8 |

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