Statistical Genomics Foundry
Paper
domain/whole-genome-duplicationtopic/ks-distribution

ksrates: positioning whole-genome duplications relative to speciation events in Ks distributions

1. Citation

2. Access note

  • Article: only the publicly visible abstract + article metadata could be read (OUP page; full PDF/HTML body is paywalled). Method-level detail below is drawn from the ksrates documentation and repo README, not the article body; the article-vs-docs boundary is flagged per statement.
  • Article license posture: all-rights-reserved. Page carries ”© The Author(s) 2021. Published by Oxford University Press. All rights reserved.” — no Creative Commons license. Therefore §7 is own-words paraphrase; no verbatim article prose is reproduced. Functional strings (parameter names, numeric values, version) are reproduced verbatim under the functional-strings exception. The docs/repo are GPL-3.0 (reproduction-permissive), but for consistency this note keeps prose in own words and quotes only short functional/load-bearing strings.

3. Thesis (1 sentence)

ksrates is a command-line tool that positions ancient whole-genome duplication (WGD) events relative to speciation events by comparing paralog and ortholog Ks distributions in one mixed plot, after correcting for synonymous-substitution-rate differences among the lineages involved so the distributions share a common scale.

4. Problem & context

  • WGDs are dated from peaks in within-species paralog Ks distributions; speciation events are dated from between-species ortholog Ks distributions. Overlaying both in a “mixed plot” is the standard way to ask whether a WGD predates or postdates a given speciation.
  • Problem the paper targets: synonymous substitution rates differ among lineages, so a Ks value for two events of the same absolute age differs depending on which lineages’ branches it was measured on. A focal species’ paralog Ks distribution is thus not directly comparable to ortholog Ks distributions involving other (differently-evolving) species, and naive superimposition in a mixed plot can mislead the phylogenetic interpretation. [article abstract + docs]
  • Rate variation cited as real: e.g. woody vs. herbaceous growth habit, plant height correlate with substitution rate. [article]

5. Method / approach

Rate-adjustment via branch decomposition using an outgroup, in the style of a relative rate test (RRT):

  • For a focal species, a diverged (target) species, and an outgroup, the total ortholog Ks between focal and diverged species is decomposed into branch-specific contributions proportional to each lineage’s substitution rate. [docs: concept page; article]
  • Rescaling rule: the diverged-species divergence is repositioned onto the focal species’ Ks scale as twice the focal branch’s contribution. Worked example from the docs (oil palm focal, rice diverged): palm–rice ortholog Ks 1.53 decomposes into palm branch 0.365 + rice branch 1.17 (palm has the lower rate); rate-adjusted divergence Ks = 2 * 0.365 = 0.73. [docs, functional numbers verbatim]
  • The adjustment shifts each speciation line in the mixed plot by an amount proportional to the estimated rate difference between the diverged lineage and the focal species. [article abstract]

Outgroups / consensus (from docs, not the article):

  • If more than one outgroup is available for a species pair, multiple rate-adjustments are performed and averaged; consensus can be the “mean among outgroups” or a single adjustment.
  • Docs advise using “at least 3 or better 4 outgroups” to buffer the weight of misleading outliers when taking the mean.
  • Config parameter max_number_outgroups caps outgroups per pair; default = 4 (without it, all possible outgroups are used).
  • Input phylogeny is a Newick tree, leaf nodes only, must contain focal_species; short names, no spaces or underscores.

Downstream: mixture modeling of paralog Ks distributions (exponential-lognormal / lognormal), anchor-Ks clustering, and a reciprocally-retained gene-family pipeline for plant WGMs are provided but are peripheral to the rate-adjustment claim. [docs]

6. Key claims / findings

  • Naive mixed plots can yield wrong phylogenetic placement of a WGD when lineages differ in rate; rate-adjustment fixes this. [article]
  • Oil palm use case: the naive plot suggests the older τ WGD is a second palm-specific WGD; after adjustment, estimates of the same event cluster together and fall into the commonly accepted order, giving the correct positioning of the two oil-palm-lineage WGDs. [article]
  • The adjustment magnitude for each speciation event is proportional to the estimated substitution-rate difference between that lineage and the focal species. [article abstract]

7. Load-bearing statements — OWN-WORDS (restrictive-license mode; functional strings verbatim)

  • Rate-adjustment purpose (paraphrase of abstract): ksrates compares paralog and ortholog Ks distributions to position ancient WGDs relative to speciation events in a phylogeny while adjusting for substitution- rate differences among the involved lineages.
  • Why single-lineage comparison fails (paraphrase): Ks estimates for events of equal absolute age differ with the lineages’ synonymous substitution rates, so a focal species’ paralog Ks distribution is not directly comparable to ortholog Ks distributions involving other species, and naive superimposition can mislead. [article]
  • Decomposition method (paraphrase of docs): using a focal species, a diverged species, and an outgroup, the raw ortholog Ks is split into branch-specific contributions (relative-rate-test style), then the divergence is rescaled to the focal Ks scale as twice the focal branch contribution.
  • Functional strings (verbatim): parameter max_number_outgroups (default 4); parameter focal_species; worked values palm–rice Ks 1.53 = 0.365 (palm) + 1.17 (rice), adjusted 2 * 0.365 = 0.73; docs wording “at least 3 or better 4 outgroups”; software v2.1.0; license GPL-3.0.

8. Stated scope, assumptions, limitations

  • Scope: Ks-based (synonymous-divergence) dating of WGDs relative to speciations; genomic or transcriptomic coding sequences. [article]
  • Rate-adjustment requires at least one outgroup per species pair to decompose branches (the RRT-style method is undefined without an outgroup). [docs — inherent to method]
  • Averaging over outgroups is a robustness device against outlier adjustments; docs frame ≥3–4 outgroups as advisable, not mandatory. [docs]

9. Failure modes / invalidity patterns

  • Rate heterogeneity among lineages is the core invalidity condition for naive mixed plots — it produces incorrect before/after-speciation placement of WGD peaks (the oil palm τ misassignment). ksrates is the correction, so the failure mode is precisely “comparing un-adjusted Ks across differently-evolving lineages.” [article]
  • Single-outgroup adjustments can be thrown off by outlier branch estimates; the docs’ remedy is averaging over several outgroups (hence the ≥3–4 advice). [docs]
  • [summarizer-inferred, not verified in source] Ks saturation at high divergence is a known general limitation of Ks methods; neither the accessible article text nor the docs pages I read state an explicit Ks upper bound (e.g. Ks<1.5) for validity of the correction. Do not attribute such a bound to this source.

10. What the source does NOT address (confident silences)

  • No hard minimum number of outgroups is stated in the article; the article speaks of “an outgroup species” (i.e. ≥1). The ≥2 or ≥3–4 figures come from the documentation’s advice, not a stated requirement, and not from the article. See §12 Q2.
  • No explicit Ks validity/saturation ceiling for the correction in the material accessed. See §12 Q4.
  • Full statistical detail of the decomposition and consensus (accessible only in the paywalled body / supplement) not captured here.

11. Open questions / ambiguities

  • Exact verbatim article wording of the plot-reading rule (WGD-before vs -after speciation) could not be read (paywall); the older/younger-Ks logic in §12 Q5 is the standard mixed-plot interpretation, not a verified verbatim quote.
  • Whether “mean among outgroups” or “single adjustment” is preferred in which situations is only lightly described in the docs.

12. Guidance answers

Q1 — What rate-adjustment corrects and HOW (branch decomposition / RRT). ANSWERED. It corrects inter-lineage synonymous-substitution-rate differences that make Ks distributions sit on different scales. Method: relative-rate-test-style branch decomposition — for focal + diverged + outgroup species, split the ortholog Ks into per-branch contributions, then rescale the divergence onto the focal species’ Ks scale as twice the focal branch’s contribution (docs example: palm–rice 1.53 → palm 0.365 + rice 1.17 → adjusted 2*0.365 = 0.73). Article states the method decomposes raw Ks into branch-specific contributions “with the help of an outgroup species, similar to relative rate testing.”

Q2 — MINIMUM OUTGROUPS: does the paper state a hard minimum? (skill asserts ”>=2”) The article states NO hard minimum. It refers to “an outgroup species” (singular) — i.e. the method needs at least one outgroup, but no explicit floor of 2 (or any number) is asserted in the accessible article text. The documentation advises “at least 3 or better 4 outgroups” to buffer outlier adjustments when averaging, and sets max_number_outgroups default = 4. So the downstream skill’s ”>=2” is NOT sourced from this paper (paper is silent on a numeric minimum) and does not match the docs’ advice either (which says 3–4). ABSENT in the paper; docs say ≥3–4 advised, not mandatory.

Q3 — Failure attributed to single-lineage Ks when comparing WGDs across differently-rated lineages. ANSWERED. Ks estimates for equally-old events differ with each lineage’s synonymous substitution rate, so a focal species’ paralog Ks distribution is not directly comparable to ortholog Ks distributions involving other species; straightforward superimposition in a mixed plot can be misleading and can misplace a WGD relative to a speciation (the oil palm τ event was wrongly read as palm-specific in the naive plot). This is the statement backing the “rate-adjustment is mandatory when rates differ” claim. [article]

Q4 — Ks validity range for correction (e.g. Ks<1.5)? ABSENT. No Ks upper bound for correction validity is stated in the accessible article text or the docs pages read. The value 1.53 appears only as an example palm–rice ortholog Ks, not as a threshold. Do not cite this source for a Ks<1.5 (or similar) cutoff.

Q5 — Reading WGD-before vs -after speciation off the mixed plot. PARTIALLY ANSWERED (interpretation captured; exact verbatim rule not accessible). Principle: in the adjusted mixed plot the paralog WGD peak(s) are compared against the rate-adjusted ortholog speciation line(s) on a common Ks axis. Standard reading (Ks increases with age): a WGD peak at older (higher) Ks than a speciation line → WGD predates that speciation → shared by both lineages; a WGD peak at younger (lower) Ks than the line → WGD postdates the speciation → lineage-specific. In the oil palm case, adjustment moved the τ event to its “commonly accepted order,” correcting the naive misplacement. [Note: an automated extraction of the docs stated a left/right rule in the opposite direction; that could not be verified and is not trusted here — the older/younger-Ks logic above is the standard interpretation, not a verified verbatim quote from the source.]

Q6 — Citation, pages, DOI, license posture. ANSWERED. Sensalari, Maere & Lohaus, Bioinformatics 38(2):530–532 (2022), DOI 10.1093/bioinformatics/ btab602. Third author = Rolf Lohaus (not Zwaenepoel) — confirmed. License: all-rights-reserved (”© The Author(s) 2021. Published by Oxford University Press. All rights reserved.”); no CC-BY. Software (repo/docs) is GPL-3.0.

(none — no project framing added; all statements sourced to article abstract/metadata or ksrates docs/repo as flagged.)