Output Formats

--format tsv|csv|json applies to both subcommands and every output mode: the pairwise dN/dS tables below, and the eskaks vcf tables further down. The examples in this first section use eskaks fasta, but the format rules (quoting, special-value normalisation) are identical for the VCF path.

TSV (default)

Tab-separated values. Standard bioinformatics format.

eskaks fasta alignment.fasta -o results
# → results_pairwise_results.tsv

Seq1    Seq2    dN  dS  dN/dS
gene_A  gene_B  0.052300    0.321400    0.162700

CSV

Comma-separated values. Compatible with Excel and pandas.

eskaks fasta alignment.fasta --format csv -o results
# → results_pairwise_results.csv

JSON

JSON array of objects. Best for programmatic parsing.

eskaks fasta alignment.fasta --format json -o results
# → results_pairwise_results.json

[
{"seq1":"gene_A","seq2":"gene_B","dN":0.052300,"dS":0.321400,"dN_dS":0.162700},
{"seq1":"gene_A","seq2":"gene_C","dN":0.089100,"dS":null,"dN_dS":null}
]

Special values (the same rule holds for every ratio: dN/dS, pN/pS, and πN/πS): - NaN (saturation, no comparable codons, or a 0/0 ratio) → null - Infinity (zero denominator over a positive numerator: dS=0 & dN>0, or pS=0/πS=0 with pN>0/πN>0) → null - -0.0 → 0.0

--format applies to every output mode, not just the default pairwise table: --lineage, --group-average, and --window-size also emit a valid JSON array (or TSV/CSV) with the fields for that mode.

In TSV/CSV, any sequence id or gene name that contains the delimiter, a quote, or a newline is quoted (RFC 4180), so the columns never shift.

Output modes

Mode	Flag	Output file
Pairwise (default)	-	<prefix>_pairwise_results.<ext>
Lineage summary	--lineage	<prefix>_lineage_summary.<ext>
Group average	--group-average	<prefix>_group_avg_dn_ds.<ext>
Sliding window	--window-size N	<prefix>_pairwise_windows.<ext>

SVG Plots (eskaks fasta)

Add --plot to generate SVG visualizations:

• Pairwise mode: dN/dS histogram (<prefix>_dnds_histogram.svg)
• Window mode: dN/dS along the alignment (<prefix>_window_plot.svg)
• Group mode: Bar chart with CI (<prefix>_group_dnds.svg)
• Lineage mode: Bar chart by lineage (<prefix>_lineage_dnds.svg)

eskaks vcf outputs

The VCF path writes one always-on table plus opt-in files per flag. All tables honour --format tsv|csv|json and the same quoting / special-value rules as above.

File	Flag	What it holds
<prefix>_pnps.<ext>	(default)	Per-gene pN/pS: counts, fractional N/S sites, the ratio, the neutrality-test p-value and BH/Bonferroni corrections, plus gene coordinates. Full column list.
<prefix>_variants.<ext>	--variants	One row per coding SNP: position, base change, S315T-style amino-acid change, AF, and effect (synonymous / missense / nonsense / stop_loss). Details.
<prefix>_diversity.<ext>	--diversity	Per-gene πN/πS, Watterson θ and Tajima's D (needs a sample size — a multi-sample VCF or --vcf-list). Details.
<prefix>_codons.<ext>	--codon-scan	Per-codon recurrence: distinct nonsynonymous alleles per residue, the codon's combinatorial opportunity, carrier bounds, and a one-sided upper-tail p/q over the whole coding genome. Details.
<prefix>_mk.<ext>	--mk	Per-gene McDonald-Kreitman 2×2 table, Neutrality Index, α and a Fisher p/q. Details.
<prefix>_pnps_manhattan.svg, <prefix>_pvalue_manhattan.svg	--plot	Manhattan plots of pN/pS and of −log10(p) by genome position, significant genes outlined.
<prefix>_report.html	--report	Self-contained interactive dashboard — no CDN or internet needed. See the live example.

JSON emits a proper array of typed objects (numbers stay numeric, NA/undefined become null, -0.0 normalises to 0.0), so the tables drop straight into pandas.read_json or jq. The --divergence <FILE> input is read, not written — it feeds the report's polymorphism-vs-divergence panel.