eskaks

Fast pairwise dN/dS and per-gene pN/pS for molecular evolution and Mycobacterium tuberculosis selection analysis. Nei-Gojobori and Li (1993) with precomputed lookup tables, 1000× faster than existing tools.

Get started Install View on GitHub

New to selection analysis?

Start with the getting-started tutorial — it runs a full analysis on the bundled example data with no background assumed. Every unfamiliar term is defined in the glossary, and hovering a dotted abbreviation like dN/dS shows its meaning.

In one line: eskaks measures natural selection on genes by comparing how fast amino-acid-changing mutations accumulate versus silent ones. A ratio below 1 means selection is removing harmful changes (a conserved gene); a ratio above 1 flags genes where change is favoured (drug targets, antigens).

What can it do?

• Pairwise dN/dS

  ---

  Codon-aligned FASTA in, dN/dS out — Nei-Gojobori or Li/LPB93, sliding windows, per-lineage and per-group summaries.

  Models

• Per-gene pN/pS

  ---

  VCF + reference + GFF3 → per-gene pN/pS, an exact neutrality test, FDR correction, an MK screen, and π / Watterson θ / Tajima's D.

  VCF analysis

• Interactive report

  ---

  A single self-contained HTML dashboard — Manhattan, volcano, QQ, MK — with a colour-blind mode and CSV/JSON/Print export. No internet needed.

  See the live report

• Fast & pipeline-ready

  ---

  Precomputed lookup tables + rayon parallelism, stdin support, JSON output, and a non-zero exit on error.

  Performance

Try it in one command

Pairwise dN/dS (FASTA)Per-gene pN/pS (VCF)No data yet?

eskaks fasta alignment.fasta --report -o results

Reads codon-aligned CDS, writes results_pairwise_results.tsv and an interactive results_report.html.

eskaks vcf --ref genome.fasta --gff genes.gff3 --vcf variants.vcf \
  --genetic-code 11 --report --variants --diversity -o scan

Writes per-gene pN/pS, a per-variant table (S315T-style keys), diversity statistics, and the report.

eskaks --demo

Runs both the dN/dS and per-gene pN/pS analyses on bundled example data - no input files needed.

How does it compare?

Feature	eskaks	KaKs_Calculator	BioPython	PAML yn00
Speed (100 seqs)	6 ms	7,703 ms	111,619 ms	697 ms
Li model R² vs KaKs	1.000	reference	—	—
Nei model R² vs KaKs	0.999	reference	0.996	—
Genetic codes	20 tables	1	1	limited
pN/pS from VCF
JSON output
Stdin pipe
Parallel	(rayon)

Citation

Ruiz-Rodriguez P, Coscollá M (2026). eskaks: fast pairwise dN/dS and per-gene pN/pS from sequences or VCFs. Zenodo. https://doi.org/10.5281/zenodo.21992154