Sampled palm civets and raccoon dogs at a Guangdong live-animal market following a January 2004 SARS case, and 1,107 additional farmed palm civets across 25 farms in 12 provinces. Found SARS-CoV-like virus in all 91 market-origin civets and 15 market raccoon dogs tested, but in none of the 1,107 farm-raised civets sampled later — a stark market-positive/farm-negative split. Phylogenetic and nonsynonymous/synonymous substitution-rate analysis showed rapid viral evolution in civets consistent with adaptation toward the human-infecting genotype, with the market population’s virus more evolutionarily advanced than an inferred ancestral “prototype” group. relevance_note: the primary source for the “market amplification” pattern explicitly invoked in the zoonosis case for SARS-CoV-2 — animals are far more likely to carry/amplify the virus once concentrated and mixed at markets than on farms, directly informing how much weight to put on Wuhan-market wildlife co-location versus farm-of-origin.
Observations
O-54 - All market palm civets and raccoon dogs at Xinyuan market tested positive for SARS-CoV-like virus
At the Xinyuan live-animal market (Guangdong, January 2004), all 91 palm civets and all 15 raccoon dogs sampled tested RT-PCR positive for a SARS-CoV-like virus, and 22 of 24 environmental swabs (cages, walls, cash tables) were positive.
Link to original
O-55 - No SARS-CoV-like virus in 1,107 farmed palm civets across 25 farms in 12 provinces
None of 1,107 farm-raised palm civets sampled across 25 farms in 12 Chinese provinces tested positive for SARS-CoV-like virus (0/1,107).
Link to original
O-56 - Same owner's civets negative on the farm but positive at the market
One farmer’s 169 palm civets held on his Henan farm all tested negative, whereas his 17 civets relocated to the Xinyuan market all tested positive - the same source animals were positive only at the market.
Link to original
O-57 - Graded SNV series in the civet spike gene from a zero-mutation prototype to human epidemic strains
Sequencing the spike gene across market civets and raccoon dogs defined 27 signature nucleotide variations (SNVs) forming a graded series: a prototype group with 0 SNVs (two raccoon dogs and one civet), market civets plus an early patient with 2-7 SNVs, and human epidemic strains with 16-27 SNVs.
Link to original
Hypotheses
H-29 - The Xinyuan market, not the farms, was where civets were infected and the virus amplified and spilled to humans
Palm civets were not infected on their farms of origin but became infected and amplified the SARS-CoV-like virus at the live-animal market, which then acted as the source of sporadic human SARS cases.
Link to original
H-30 - Palm civets are an amplifying intermediate host, not the natural reservoir of SARS-CoV
Palm civets act as a highly susceptible amplifying/intermediate host for SARS-CoV at markets rather than being its natural reservoir, which remains to be identified.
Link to original
Arguments
A-56 - Market-positive-farm-negative split with a within-owner control locates infection at the market
The 100%-positive market vs 0/1,107 farm contrast, and especially the same owner’s civets being negative on the farm yet positive at the market, force the inference that infection occurred at the market rather than on the farms or being intrinsic to the animals.
Reasoning
The within-owner comparison controls for the animals’ genetic stock, husbandry, and farm of origin: the 169 farm civets and the 17 market civets came from the same owner/source, so a difference in infection status cannot be attributed to which animals they were. The variable that changed is location (farm vs market). Since the farm animals were uniformly negative (0/169, and 0/1,107 across 25 farms in 12 provinces) while the market animals were uniformly positive (17/17, and 91/91 civets plus 15/15 raccoon dogs, with 22/24 environmental swabs positive), the market environment - dense mixing of many individuals and species over shared cages and surfaces - is where transmission and amplification occurred. This rules out the farms as the site of civet infection and supports the market as the amplification and spillover locus. The design does not by itself identify the ultimate reservoir that seeded the market.
Verdict — approved (checked)
Reconstruction. Premises: (i) farm and market civets from the same owner shared genetic stock/husbandry/origin; (ii) they were uniformly negative on the farm and uniformly positive at the market; (iii) the broader farm sample was 0/1,107 and the market sample uniformly positive incl. environment. Load-bearing step: a controlled contrast holding the animals constant and varying only location assigns the difference in infection status to location — i.e. a differential-comparison (matched-control) inference.
Evaluation. Assuming the premises, the step goes through for the two claims that carry weight: infection is not intrinsic to the animals (the same stock was negative earlier) and did not occur on the farms. The one surviving undercutting defeater is that infection could have happened during transport/trade between farm and market rather than physically inside the market building; this does not deny the premises. But it does not rescue the farm or the animal-intrinsic alternative, and it collapses the “market vs farm” dichotomy the argument actually tests into “market/trade chain vs farm” — immaterial to what the argument feeds (H-29’s market-amplification claim). The argument itself disclaims identifying the ultimate reservoir. The step is real; approved. Traced directly from the tabulated counts, author-blind.
Link to original
A-57 - Prototype-to-epidemic SNV progression indicates active viral adaptation in civets at the market
A graded SNV series whose zero-mutation prototype is found in market animals and which progresses toward the human epidemic genotype indicates the virus was actively evolving within civets at the market, consistent with market amplification/adaptation rather than a long-established stable civet reservoir.
Reasoning
The 27 SNVs partition sequences into an ordered ladder (0 → 2-7 → 16-22 → 25-27 mutations). The least-evolved prototype (0 SNVs) is recovered from market animals, and intermediate genotypes (2-7 SNVs) are shared between market civets and an early human patient, while fully human-epidemic strains carry the most mutations. An ordered accumulation of substitutions within the market population, with intermediates linking animal and human sequences, is the signature of ongoing microevolution/adaptation in the host being sampled - here civets at the market. A stable long-term reservoir would not be expected to yield a graded, still-diversifying series culminating in the human-infecting genotype. This supports civets being an amplifying host in which the virus adapts toward human infectivity, and that this was occurring at the market that seeded human cases.
Verdict — approved (checked)
Reconstruction. Premises: the 27 SNVs sort sequences into an ordered ladder (0 → 2-7 → 16-22 → 25-27); the 0-SNV prototype is in market animals; 2-7-SNV intermediates are shared between market civets and an early human case; the most-mutated genotypes are the human epidemic strains. Load-bearing step: an ordered, still-diversifying substitution series sampled within a host population, polarised from an animal-borne least-derived form toward the human genotype, is the signature of ongoing microevolution in that population rather than sampling from a fixed long-established reservoir.
Evaluation. Conditional on the premises the step holds for the hedged conclusion actually stated (“indicates”/“consistent with”, not “proves”). The polarity — least-derived form in animals, most-derived in humans — establishes the animal→human direction, and the presence of multiple co-circulating market-civet genotypes shows active diversification in that population, which a stable long-term reservoir (expected to present standing diversity without a graded ladder culminating in the human strain being freshly sampled) would not produce. Candidate undercutting defeater: the ladder could reflect a transmission chain whose mutations partly accumulated in humans rather than within civets. This trims the strongest reading (“adapting within civets specifically”) but not the stated conclusion that civets were an amplifying host at the market rather than the stable reservoir — which the animal-side prototype and shared intermediates carry regardless. Approved; traced from the SNV ladder, author-blind.
Link to original