A suite of regression-based tools for Ribosome profiling data analysis
install.packages('BiocManager')
BiocManager::install("Goodarzilab/Ribolog")Run the code below in your terminal to build and activate a conda environment with all the dependencies of Ribolog inside it.
conda env create -f 'https://raw.githubusercontent.com/goodarzilab/Ribolog/master/environment.yml'
conda activate ribolog # Now you're inside the conda environment
R -e "BiocManager::install('Goodarzilab/Ribolog', dependencies = FALSE)"Please refer to the "Ribolog_Full_Tutorial.ipynb" file for a full tutorial of the package.
Converts aligned, sorted, and indexed BAM files into per-read P-site assignments: read-length distributions,
P-site offset detection, and reading-frame/periodicity QC. Produces the reads_psite_list object used by
later modules.
Corrects RPF counts for codon-level stalling bias (CELP: Consistent Excess of Loess Preds) and analyzes ribosome dwell times per codon/amino acid, then normalizes and filters RNA/RPF read counts into the transcript-by-sample count matrices used by Modules 3-6.
Three tools for assessing replicate reproducibility and biological signal before testing: PCA of translational efficiency, the proportion of null (non-differentially-translated) features, and correlograms of equivalent replicate-vs-replicate TER tests.
Tests the size and significance of differential translational efficiency between biological samples via logistic regression. Although better results are always obtained with sufficient replicates, Ribolog is able to perform the TER test with only one replicate per sample. The TER test is not restricted to pairwise comparisons; any number of samples described by several attributes (covariates) can be compared in a single model.
Two advanced statistical tools for TER testing: 1) empirical null hypothesis testing, which derives the null distribution directly from replicate-vs-replicate comparisons in your own data instead of assuming a theoretical one, reducing false positives from batch effects, mapping noise, or overdispersion. 2) meta-analysis, which combines correlated test results (e.g. from separate datasets, or rep-by-rep sub-tests) into a single consensus effect size and p-value.
Tests differential usage of upstream ORFs and stop-codon readthrough by comparing the distribution of P-site reads across the 5'UTR, CDS, and 3'UTR regions of each transcript between biological samples.
Ribolog was developed by Hossein Asgharian and Sohit Miglani at UCSF supervised by Hani Goodarzi and Adam Olshen. More modules are being prepared and will be released in near future.
For questions and comments, email us at: