Skip to content

4.7 Codon-level analysis

Codon-level analysis summarizes ribosome profiling signal at codon resolution. These modules are useful for studying codon pausing, codon occupancy, codon decoding time, codon selection time, coverage-dependent variation, local meta-codon profiles, and differential codon enrichment between experimental groups.

Required upstream results

Before running codon-level analysis, make sure that the upstream Ribo-seq results have passed the basic quality-control steps:

  • A reliable offset table has been generated and used for RPF density construction.
  • The selected reads show clear 3-nt periodicity.
  • CDS enrichment is reasonable.
  • The merged RPF density table has enough reads for the selected genes.
  • Biological replicates show acceptable reproducibility.

Most codon-level modules use the merged density table generated by rpf_Merge, for example:

RIBO_merged.txt

For CDT and CST analysis, a matched RNA-seq density table is also required:

RNA_merged.txt

Submodules

Section Command Purpose
4.7.1 codon pausing scores rpf_Pausing Calculate relative codon pausing scores from codon-level RPF density.
4.7.2 codon occupancy rpf_Occupancy Calculate absolute and relative codon occupancy.
4.7.3 codon decoding time rpf_CDT Estimate codon decoding time by combining Ribo-seq and RNA-seq density.
4.7.4 codon selection time rpf_CST Estimate codon selection time using iterative Ribo-seq/RNA-seq codon-level calculations.
4.7.5 coefficient of variation rpf_CoV Quantify coefficient of variation across CDS regions.
4.7.6 cumulative CoV rpf_Cumulative_CoV Quantify cumulative coefficient of variation along the CDS.
4.7.7 codon metaplot rpf_Meta_Codon Extract and plot local RPF density around selected codons or codon motifs.
4.7.8 codon odd ratio rpf_Odd_Ratio Compare codon-associated RPF enrichment between control and treatment samples.
4.7.9 codon pause site rpf_PSplot Visualize the detected local pause sites on the raw RPF profiles.

Suggested directory structure

The codon-level modules are organized under sce/4.ribo-seq/5.riboparser/ (see the project layout in 3 New project). Sub-directories 01.qc to 10.quantification are created by the quality-control and gene-level modules; create the codon-level ones (11–19, matching the 4.7.1–4.7.9 sections above) as follows:

mkdir -p 11.pausing_score 12.codon_occupancy 13.codon_decoding_time \
         14.codon_selection_time 15.coefficient_of_variation 16.cumulative_of_cov \
         17.meta_codon 18.odd_ratio 19.pause_site_plot

Input dependency

Most codon-level modules consume the merged RPF density matrix produced by rpf_Merge in 05.merge (see 4.5.5 Merge density); only rpf_CDT and rpf_CST additionally require a matched merged RNA density matrix from the RNA-seq branch. The per-module input and output are summarized below:

Section Main input Main output
Pausing merged RPF density (-r) relative codon pausing scores; per-sample tables and figures
Occupancy merged RPF density (-r) absolute and relative codon occupancy tables
CDT merged RPF density (--rpf) + merged RNA density (--rna) codon decoding-time tables, sample correlation, heatmaps, and rank plots
CST merged RPF density + merged RNA density codon selection-time tables and figures
CoV merged RPF density (-r); optional sample-group table (-g) transcript CoV tables, group comparison/fitting tables, and plots
Cumulative CoV merged RPF density (-r); optional transcript list (-l) meta cumulative-CoV tables, meta/support curves, and per-transcript figures
Codon metaplot merged RPF density (-r); optional codon/motif list (-c) per-motif meta density and sequence-context tables and line plots
Odd ratio merged RPF density (-r) + control and treatment sample names (-c/-t) site-level local-pause table, codon-level summary, and figures
Pause-site plot rpf_Odd_Ratio site table (-i) + merged RPF density (-r); optional target gene list (--target-list) per-gene, per-sample PSplot figures; optional pause-metrics table (--export-metrics)

Note that rpf_PSplot depends on the site-level table written by rpf_Odd_Ratio, so run section 4.7.8 before 4.7.9. For modules that support two input formats, both the compact JSONL/JSONL.GZ file (for example sce_rpf_merged.jsonl.gz) and the plain TXT table are accepted.

Notes

  • Use the same offset and frame-selection strategy across modules when results need to be compared.
  • For high-quality Ribo-seq data, frame 0 is commonly used after offset correction, but all can be useful during exploratory analysis.
  • Parameters such as --tis, --tts, and --stop help avoid start/stop-proximal artifacts.
  • Codon-level statistics are sensitive to low coverage, so choose -m carefully according to sequencing depth and study design.