Because deliberate engineering (given early-2020 knowledge) would have started from a documented published virus backbone, and SARS-CoV-2’s backbone matches none, construction from a known reverse-genetics system is not supported by the genetic data.
Reasoning
Laboratory generation of a coronavirus via reverse genetics typically begins from a previously characterised infectious-clone backbone. If SARS-CoV-2 were such a construct, its non-spike genome should betray close identity to one of those documented systems. The observed backbone instead matches no previously used platform. Under the premise that engineers would build on an existing, published backbone rather than an unknown/unpublished one, the absence of any match is taken as evidence against a laboratory construct. The premise (that a known backbone would be used) is the load-bearing assumption; an unpublished platform is not formally excluded by the data.
Verdict — approved (checked)
Reconstruction. Premises: (i) reverse-genetics generation of a coronavirus begins from a previously characterised infectious-clone backbone; (ii) such a construct would retain close non-spike identity to its source backbone; (iii) SARS-CoV-2’s backbone matches no previously described system; (iv, load-bearing) engineers would build on an existing, published backbone. Load-bearing step: modus tollens — if built from a known published backbone we would see a match; no match; therefore not built from a known published backbone.
Evaluation. Conditional on the premises the step is a clean modus tollens and goes through. The natural undercutting move — that an unknown/unpublished backbone could have been used — does not deny the premises but is already absorbed by the conclusion’s scope: the statement claims only that construction from a known reverse-genetics system is unsupported, not that engineering is excluded. The argument does not overreach beyond what the data license (it explicitly flags the unpublished-platform gap). No defeater survives against the scoped claim. Approved; the inference is elementary and traced author-blind. Whether premise (iv) is actually true is a truth question priced at step 7-8, not here.