Tettelin et al. 2008 — Comparative genomics: the bacterial pan-genome

1. Citation

Tettelin H, Riley D, Cattuto C, Medini D. 2008. “Comparative genomics: the bacterial pan-genome.” Current Opinion in Microbiology 11(5):472–477. doi:10.1016/j.mib.2008.09.006. PMID: 19086349. Publication types: Comparative Study; Review. MeSH: Bacteria/genetics; Genetic Variation; Genome, Bacterial; Genomics/methods; Synteny.

Publisher: Elsevier — all-rights-reserved (no open-access/CC statement found).

2. Access note — READ THIS FIRST (hard paywall boundary)

I could only read the abstract + PubMed metadata. The full text is behind the Elsevier paywall. All full-text routes tried failed:

DOI → linkinghub.elsevier.com: redirect stub only, no body.
ScienceDirect article page: HTTP 403.
ResearchGate PDF page: HTTP 403.
Semantic Scholar: blank body.

Everything the guidance most wants — the exact growth-law functional form, the exact exponent symbol, the open/closed inequality and its boundary value, any minimum-genome count, the named open/closed species, and the exact gene-class vocabulary — lives in the paywalled body text and figures, NOT in the accessible abstract. None of it was recovered from the source. Do not treat this note as recovering those functional strings; it recovers the abstract only.

I deliberately did not reconstruct the equations from secondary/citing papers. Secondary sources disagree with each other on both the symbol and the sign/boundary convention (see §11), so importing them would be unfaithful and would encode a possibly-wrong functional string. This note therefore leaves those items OPEN.

3. Thesis (from abstract, own words)

As multiple genomes per species become cheap and routine, pan-genome analysis — characterizing the size of the gene repertoire accessible to a species together with an estimate of how many whole-genome sequences are needed for proper analysis — is increasingly applied; existing models vary in accuracy and applicability, and the authors propose a new model of broad applicability with worked examples.

4. Problem & context (from abstract)

Bacterial genome sequencing became easy/accessible → multiple strains per species are being sequenced.
These multi-strain sets enable in-depth analysis of intra-species diversity.
Pan-genome analysis characterizes the accessible gene repertoire AND estimates the number of whole-genome sequences required for proper analysis.
“Different models exist for the analysis and their accuracy and applicability depend on the case at hand.” (abstract, paraphrased — this is the paper’s stated motivation for a new model.)

5. Method / approach (only what the abstract states)

The paper is a review that (a) discusses current pan-genome models and (b) suggests a new model of broad applicability, with examples of its implementation.
The specific mathematical form of the “current models” and of the “new model,” the fitting procedure, and the worked examples are all in the paywalled body — not accessible here.

6. Key claims / findings (verified subset only)

Pan-genome analysis simultaneously (i) characterizes the size of the species-accessible gene repertoire and (ii) estimates the number of whole-genome sequences required for proper analysis. [abstract — verified]
Multiple existing models exist; accuracy/applicability is case-dependent. [abstract — verified]
The authors propose one new, broadly-applicable model, demonstrated on examples. [abstract — verified]
Quantitative results (exponent values, species classifications, genome-count thresholds): not accessible — paywalled.

7. Load-bearing statements — mode: own-words paraphrase (restrictive license)

License is all-rights-reserved → no verbatim prose reproduced. The single load-bearing accessible statement (abstract) is paraphrased in §3–§6. The abstract contains no functional strings (no equations, exponents, thresholds, or species names), so the functional-strings exception yields nothing to quote here. The functional strings the guidance requires are in the paywalled body and were not read.

8. Stated scope, assumptions, limitations (from abstract)

The authors themselves frame model choice as case-dependent: “their accuracy and applicability depend on the case at hand” (paraphrased). I.e., no single existing model is claimed universally correct — the motivation for their new “broadly applicable” model.
Further limitations stated in the body are not accessible.

9. Failure modes / invalidity patterns

Not recoverable from the abstract. The abstract implies (but does not spell out) that applying a mispecified model to the wrong case gives poor accuracy [summarizer-inferred from “accuracy and applicability depend on the case at hand”]. Any named diagnostics, value ranges, or invalidity criteria are in the paywalled body — not read.

10. What the source (accessible portion) does NOT address

Everything below is absent from the accessible abstract; presence in the full text is unknown/likely but unverified:

The power-law / Heaps’-law functional form and the exponent symbol.
The open-vs-closed decision inequality and its boundary value.
Any minimum number of genomes required to estimate the exponent.
Named open vs closed species.
The gene partition class vocabulary (core / dispensable / accessory / unique) precise definitions.

11. Open questions / ambiguities left unresolved

Unresolved BY THIS NOTE due to the paywall (all five guidance functional-string targets):

Growth-law form + exponent symbol. Citing papers report incompatible conventions, so this cannot be settled from them: some write Heaps’ law as pan-genome size n = κ · N^γ (open if the exponent > 0 and the curve is non-asymptotic; one source phrased “closed” as exponent < 0); others frame it on new genes discovered decaying as ∝ N^(−α) with open if α ≤ 1, closed if α > 1; still others state “open if the exponent < 1, closed if > 1.” These are NOT the same statement, and the boundary value (0 vs 1) and the symbol (α vs γ) both differ. [The above are secondary-source claims, explicitly NOT from Tettelin 2008; recorded only to show why the item must be verified against the paywalled original, not to answer it.]
Whether the rule is stated on the raw exponent or on (1−exponent): unknown from accessible text.
Minimum-genome count: unknown (the abstract confirms the paper estimates a required number, but gives no figure).
Named species: unknown from accessible text.
Exact class vocabulary in THIS paper: unknown from accessible text.

To close these, the paywalled full text (or a licensed PDF) must be read directly. Do not backfill from memory or from citing papers.

12. Guidance answers (targeted questions)

Power-law/Heaps’-law form + exact exponent symbol (verbatim required): NOT ANSWERED from the source — paywalled. The functional form and exponent symbol are in the body/figures, not the abstract. Not recovered. (Secondary-source conventions conflict; see §11 — deliberately not adopted.)
Open vs closed decision rule + exact inequality and boundary (verbatim required): NOT ANSWERED from the source — paywalled. Whether the rule is on α or (1−α), and boundary 0 vs 1, is unresolved here. Not recovered.
Minimum number of genomes to estimate the exponent reliably: NOT ANSWERED with a number. The abstract confirms the analysis includes “an estimate of the number of whole genome sequences required for proper analysis,” but states no threshold value. Any specific minimum is in the paywalled body — not read.
Named open vs closed species (verbatim): NOT ANSWERED from the source — paywalled. No species are named in the abstract.
Core / dispensable / finer partition vocabulary (exact classes): NOT ANSWERED from the source — paywalled. The abstract does not use these class terms.

Must-extract functional strings (growth-law form + exponent symbol; open/closed inequality + boundary; minimum genome count; named species): all unrecovered — behind the Elsevier paywall. This source note is a low-recoverability stub pending licensed full-text access.