Review Papers Claim by Claim

Visual-first review — "Assessing the Epidemic Potential of RNA and DNA Viruses" (Woolhouse et al., 2016)

Visual summary (key conceptual figures)

• Pathogen pyramid & R0 mapping: the paper maps four pyramid levels to R0 categories: L1 (exposure, no infection), L2 (infection, R0≈0), L3 (limited transmission, 0<R0<1) and L4 (sustained transmission, R0>1) — this link to R0 is the paper's central organizing insight and useful for surveillance triage. Pathogen pyramid mapped to R0
• Use of phylogenetics and fast substitution rates: the authors correctly note that high substitution rates in many RNA viruses (≈1–5×10^{-3} substitutions/site/year) enable near-real-time phylogenetics to test for chains of human-to-human transmission — the logic is sound and prescient given later genomic surveillance successes. Phylogenetics & substitution rates

Critical appraisal — strengths (visual bullets)

• Clear, mechanistic framing: linking R0 to pyramid levels makes the problem operationally tractable for surveillance and risk communication. R0 operationalization
• Broad taxonomic sweep: the review documents that human epidemic potential is phylogenetically widespread (22 families contain level-4 viruses), an evidence-backed caution against narrow taxonomic targeting. Taxonomic breadth of level-4 viruses
• Integrative examples: the paper synthesizes epidemiology, evolution, vector ecology (e.g., Aedes linkage to arboviral L4 status), and case histories (Ebola 2014, chikungunya vector adaptation, Zika expansion), which make arguments concrete. Case studies illustrating level transitions

Critical appraisal — limitations, blindspots, and suggestions

1. Descriptive vs. quantitative: the paper is primarily conceptual and literature-synthesis; it does not present a transparent, reproducible dataset or code for the systematic review used to assign viruses to levels, which impedes reproduction and sensitivity analyses (discovery bias, reporting bias). Evidence: the paper refers to an online technical appendix for methods but provides limited raw datasets and no open code. Limited reproducible data/code
2. Discovery & sampling bias under-addressed: while the authors mention that discovery effort influences perceived nearest relatives, the paper lacks formal quantification of how discovery biases (e.g., differing surveillance effort in bats, rodents, domestic animals) change posterior risk assessments for L2→L3 or L3→L4 transitions. Subsequent phylogenetic work (e.g., later large-scale analyses) shows this is a major effect. Discovery bias acknowledged but not quantified
3. Operational surveillance guidance is qualitative: the paper recommends prioritizing level-3 viruses for surveillance because small changes in R0 matter, but it does not supply a reproducible scoring/weighting system (e.g., incorporating reservoir prevalence, human exposure frequency, viral substitution rate) to rank the 37 level-3 viruses quantitatively. That limits actionable prioritization under constrained resources. Lack of quantitative prioritization
4. Evolution vs ecology nuance: the authors correctly contrast preadaptation (off-the-shelf) vs adaptation (tailor-made) emergence, but empirical evidence to differentiate which viruses are likely to follow which route is sparse; the paper speculates (e.g., retroviruses may be more 'tailor-made') without systematic comparative data. That leaves a gap for predictive modelling. Speculation on emergence modes lacks systematic test
5. Vector complexity simplified: the review correctly notes Aedes anthropophily is linked to arboviral L4 status, but the vector-host-virus interaction space is compressed; for instance, vector competence, vector abundance, host availability, and virus-vector genetic compatibility interact nonlinearly — a formal eco-evolutionary transmission model would improve predictive power. Vector ecology discussed qualitatively

Concrete recommended improvements (for next version / operationalization)

• Publish the full virus-level dataset (assignments to L1–L4) and underlying evidence (case counts, phylogenies, genomic accessions) with code to allow reclassification sensitivity checks.
• Implement a formal probabilistic risk model that integrates: reservoir prevalence, human exposure frequency, viral substitution rate (evolvability), tissue tropism constraints, and vector/host ecology to output posterior probability for L2→L3 and L3→L4 transitions per virus.
• Quantify discovery bias: use subsampling/rarefaction and predictive models (e.g., time-varying discovery) to show how incompleteness changes the nearest-known-relative inference and thus the assessed risk.
• Provide a reproducible surveillance-priority score (open methodology) so public-health agencies can rank the 37 level-3 viruses under budget constraints.

Short pathway to falsify key claims

• Claim: "Most human-transmissible viruses already possess transmission capability at first human infection (off-the-shelf)." Falsifier: show multiple robust phylogenetic/time-resolved reconstructions where transmissible lineages arose via within-human adaptive substitutions post-first-human-infection across diverse taxa (with temporally dense sequences demonstrating de novo adaptation and R0 rise).
• Claim: "Level-3 viruses are priorities because small changes in R0 can rapidly change threat." Falsifier: Demonstrate, with empirical outbreak data and ecological parameter estimates, that for the majority of the 37 level-3 viruses, ecological constraints (e.g., vector absence, essential tissue tropism) prevent sustained R0 increase even under plausible mutational scenarios.

Confidence, provenance, and key citation

This review and critique are grounded in the Woolhouse et al. (2016) EID paper and the authors' referenced conceptual literature; claims about R0 mathematics and phylogenetic utility are supported in the paper and by subsequent genomic epidemiology practice. Woolhouse et al. 2016 (EID)

Key primary citation used in this review:

Notes on epistemic humility: Woolhouse et al. produce a logically coherent and useful conceptual framework; however, because the paper synthesizes diverse published evidence rather than releasing a fully reproducible primary dataset and formal probabilistic models, its operational recommendations should be treated as guidance that requires quantitative, reproducible follow-up analyses before being used for high-stakes surveillance prioritization.