John Murphy, M.D., M.P.H., D.P.H., COVID-19 Long-haul Foundation
Summary
The origin of SARS-CoV-2 remains one of the most consequential unresolved questions in modern infectious-disease science. The central issue is not whether SARS-CoV-2 can be shown, from presently available evidence, to have been deliberately engineered. Rather, it is whether the genomic architecture of the virus, the biology of its spike protein, the history of coronavirus research before 2020, the availability of closely related viruses to research laboratories, and the circumstances surrounding the emergence of the pandemic together constitute a scientifically credible laboratory-associated hypothesis.
Several observations have made that hypothesis worthy of sustained investigation. SARS-CoV-2 possesses a distinctive polybasic furin-cleavage site at the S1/S2 boundary of spike created by a 12-nucleotide insertion. The insertion produces the amino-acid sequence PRRA and introduces a cleavage motif with important consequences for viral entry and tropism. The inserted sequence also contains two consecutive CGG arginine codons, a nucleotide configuration that has attracted particular attention because of its relative rarity among sarbecoviruses. At the same time, the virus possesses a receptor-binding domain capable of efficient interaction with human ACE2 and an overall genomic architecture situated within a research field that, before the pandemic, was actively exploring bat sarbecoviruses, spike-mediated host range, recombination and viral adaptation.
These observations do not constitute a molecular fingerprint proving laboratory construction. They do, however, define a constellation of biological and historical features that makes laboratory-associated emergence a legitimate scientific hypothesis. The appropriate scientific objective is therefore not to convert circumstantial evidence into certainty, but to determine whether the missing experimental, genomic and laboratory records can discriminate among competing origin pathways.
Introduction: The Question That Remains
The origin of SARS-CoV-2 is not merely an historical question. It is a question about the governance of high-consequence pathogen research, biosafety, surveillance, experimental evolution and pandemic prevention.
The virus emerged in Wuhan in late 2019 and rapidly demonstrated efficient human-to-human transmission. Genomic sequencing established that SARS-CoV-2 belonged to the sarbecovirus lineage of betacoronaviruses and was related to viruses previously identified in bats. The early genomic literature immediately identified two unusual features of the viral spike: its receptor-binding domain and a polybasic cleavage site at the S1/S2 boundary.[1]
The laboratory hypothesis subsequently developed along several related but distinct pathways. A laboratory-associated event could involve acquisition of a naturally occurring virus followed by laboratory passage or adaptation; manipulation of a naturally occurring coronavirus; recombination among coronavirus sequences; or, in the strongest version of the hypothesis, deliberate construction of a novel virus.
These possibilities should not be treated as synonymous.
The scientifically most productive question is therefore:
Could SARS-CoV-2 have emerged through research involving naturally occurring sarbecoviruses, laboratory passage, recombination, adaptation or reverse-genetics manipulation?
That question can be approached through molecular genetics, evolutionary biology, experimental virology and reconstruction of the pre-pandemic coronavirus research environment.
1. The Furin-Cleavage Site
The most conspicuous molecular feature supporting laboratory-origin inquiry is the polybasic furin-cleavage site in SARS-CoV-2 spike.
Spike is synthesized as a precursor that is subsequently processed into S1 and S2 subunits. SARS-CoV-2 contains a polybasic sequence at the S1/S2 junction that can be recognized by furin and related host proteases. The sequence is generated by a 12-nucleotide insertion and results in the amino-acid motif RRAR within the cleavage region.[1]
This feature matters because cleavage of coronavirus spike can influence entry into host cells, tissue tropism, host range and replication.
The evolutionary significance of the insertion is therefore greater than that of an ordinary neutral mutation. It occurs at a biologically consequential position in the principal protein governing viral entry.
The laboratory hypothesis does not require the assertion that a furin-cleavage site cannot evolve naturally. Rather, the relevant observation is that SARS-CoV-2 appeared with a functionally important polybasic cleavage site embedded within a virus otherwise closely related to sarbecoviruses lacking that precise configuration.
Consequently, the site became an obvious molecular feature requiring reconstruction of its evolutionary history.
2. The PRRA Insertion and Its Nucleotide Composition
The amino-acid consequence of the insertion is conventionally described as PRRA. At the nucleotide level, however, the sequence contains an additional feature that has received disproportionate attention: two consecutive CGG codons encoding arginine.
Codon usage is not deterministic. Nevertheless, the nucleotide composition of the insertion is potentially informative when reconstructing the mechanism by which the sequence arose.
The significance of the CGG-CGG observation is not that CGG is impossible in naturally occurring coronaviruses. It is that a rare-looking nucleotide configuration occurs within a newly introduced sequence precisely at a biologically important functional site.
Thus, the question is not:
“Can nature produce CGG?”
It plainly can.
The more consequential question is:
What evolutionary or molecular mechanism most plausibly produced this particular nucleotide sequence at this particular location?
Answering that question requires comparison against comprehensive sarbecovirus sequence databases, experimentally measured mutation and recombination processes, codon usage across relevant coronavirus lineages, and the sequence inventories available to laboratories before the pandemic.
3. Spike as an Experimental Target
The laboratory hypothesis gains additional significance from the central role of spike in coronavirus research.
Before SARS-CoV-2 emerged, investigators had already established that coronavirus spike proteins determine major aspects of host-cell recognition and entry. Bat SARS-like coronaviruses were being investigated precisely because their spike proteins could reveal mechanisms governing cross-species transmission.
Experimental coronavirus research therefore had a natural focus on:
- receptor-binding domains;
- ACE2 interaction;
- spike cleavage;
- host range;
- viral entry;
- recombination;
- chimeric viruses;
- and adaptation to experimentally relevant host systems.
The laboratory-origin hypothesis consequently does not require a hypothetical technology. Reverse-genetics systems capable of manipulating coronavirus genomes already existed.
Published research had demonstrated that coronavirus spike proteins could be exchanged, modified and experimentally evaluated for effects on host-cell entry and host range. These methods created a technological environment in which experimentally altering coronavirus phenotypes was scientifically feasible before the emergence of COVID-19.
The relevant question is therefore one of provenance: which experimental coronavirus systems existed, which viruses were available, what sequences were possessed, what experiments were performed, and which records survive?
4. The ACE2-Receptor-Binding Domain
A second major feature of SARS-CoV-2 is its receptor-binding domain (RBD).
The RBD interacts with angiotensin-converting enzyme 2 (ACE2), the principal cellular receptor exploited by SARS-CoV-2. Structural and biochemical studies established that the SARS-CoV-2 RBD is capable of productive interaction with human ACE2.[1]
The significance of the RBD for the laboratory hypothesis is not simply that SARS-CoV-2 binds ACE2. Many viruses evolve receptor-binding capacity.
Its importance lies in the combination of:
a highly functional human receptor-binding domain + a distinctive spike-cleavage configuration + emergence within a region containing extensive pre-pandemic sarbecovirus research.
A laboratory-development model does not necessarily require scientists to have designed every feature independently. A more biologically realistic possibility is that a naturally occurring viral backbone could have undergone experimental manipulation, recombination or adaptation involving spike sequences.
This distinction is crucial. Laboratory-associated emergence need not mean that the entire viral genome was synthetically designed from the beginning.
5. Coronavirus Recombination Provides a Mechanistic Framework
Coronavirus genomes are unusually capable of recombination. RNA-dependent RNA polymerase template switching provides a mechanism through which related coronavirus genomes can exchange genetic material.
This biological property is important when considering laboratory hypotheses because experimental coronavirus research can bring together viral sequences that might rarely encounter one another in nature.
A recombinant-origin scenario therefore does not require the artificial synthesis of an entire genome. It could theoretically involve the manipulation or propagation of naturally occurring viral sequences followed by recombination.
The resulting virus could remain overwhelmingly natural in sequence while possessing a small number of experimentally introduced or selected characteristics.
That possibility makes the distinction between “natural virus” and “engineered virus” scientifically inadequate. The relevant spectrum includes natural evolution, laboratory passage, experimental recombination, directed adaptation and synthetic construction.
6. The Research Environment Before the Pandemic
Perhaps the most important contextual element is the pre-pandemic scientific environment.
By 2019, researchers were actively investigating bat sarbecoviruses because of the recognized possibility that animal coronaviruses could cross species barriers. Studies had examined coronavirus receptor usage, spike proteins, chimeric viruses and mechanisms by which bat viruses might acquire the ability to infect human cells.
The existence of such research does not itself identify the origin of SARS-CoV-2.
It does establish, however, that the biological processes required by a laboratory-associated hypothesis were already within the scope of contemporary experimental virology.
The scientific reconstruction therefore requires more than examination of the final SARS-CoV-2 sequence. It requires reconstruction of the pre-2019 viral inventory and experimental history.
That includes:
- all sarbecovirus sequences held by relevant laboratories;
- field specimens from which those sequences were derived;
- laboratory notebooks and sequencing records;
- passage histories;
- viral culture records;
- reverse-genetics constructs;
- plasmid and clone inventories;
- sequencing raw data;
- experimental protocols;
- records of animal experiments;
- biosafety incidents and exposure reports; and
- unpublished viral genomes.
Without these data, molecular phylogenetics alone cannot reconstruct every possible laboratory pathway.
7. The Importance of the Missing Immediate Progenitor
One of the central unresolved genomic questions is the identity of the immediate evolutionary predecessor of SARS-CoV-2.
RaTG13, sampled from a bat in Yunnan and subsequently sequenced, is approximately 96% identical to SARS-CoV-2 across the genome. It is therefore closely related but not the immediate progenitor.[1]
The evolutionary distance is substantial enough that RaTG13 cannot simply be described as the precursor of SARS-CoV-2.
Subsequent discoveries of additional sarbecoviruses, including viruses sampled in Southeast Asia, have expanded the known diversity and provided sequences containing important similarities to SARS-CoV-2. These discoveries demonstrate that the natural sarbecovirus reservoir is considerably more diverse than was appreciated at the beginning of the pandemic.
At the same time, the precise virus immediately ancestral to SARS-CoV-2 has not been identified.
This creates an important evidentiary gap.
If the immediate progenitor remains absent from the natural-virus record, then a complete reconstruction of SARS-CoV-2 evolution remains incomplete.
The laboratory hypothesis therefore predicts that historical research collections could be particularly informative.
A virus need not have been publicly known to have existed in a laboratory. A specimen collected during fieldwork may have been stored, cultured, sequenced incompletely or not sequenced at all.
Consequently, the question of what viruses were physically present in research collections before December 2019 is more important than simply asking which genomes had been published.
8. The Wuhan Geographic Convergence
The geographic dimension cannot establish causality, but it is relevant to hypothesis generation.
The first recognized outbreak occurred in Wuhan, while Wuhan simultaneously housed a major research program devoted to bat-associated coronaviruses.
This convergence is scientifically unusual enough to justify investigation.
The laboratory hypothesis therefore rests partly on a simple epidemiological principle:
when an unusual pathogen first appears in a city containing an institution conducting research on closely related pathogens, the laboratory should be included in the epidemiological investigation.
That does not determine the answer.
It determines where investigators should look for evidence.
9. From “Man-Made” to “Laboratory-Associated”
The phrase “man-made virus” is scientifically imprecise.
There are at least four distinct hypotheses:
Hypothesis A — Natural zoonotic emergence
A virus circulating in wildlife crosses into humans without laboratory involvement.
Hypothesis B — Laboratory acquisition and accidental release
A naturally occurring virus is collected, cultured or otherwise handled in a laboratory and subsequently escapes through an accidental exposure.
Hypothesis C — Laboratory adaptation or manipulation
A naturally occurring virus undergoes laboratory passage, recombination, adaptation or genetic modification before accidental release.
Hypothesis D — Purposeful synthetic construction
A virus is deliberately designed or assembled using synthetic or reverse-genetics techniques.
The molecular observations surrounding SARS-CoV-2 are most directly relevant to hypotheses B and C.
The evidence required to establish hypothesis D would be substantially more specific: identifiable engineering signatures, documented precursor constructs, laboratory records or a reproducible reconstruction demonstrating a particular engineering pathway.
Therefore, a rigorous scientific discussion should avoid treating the laboratory-associated hypothesis as synonymous with deliberate bioweapon construction.
10. The “Constellation” Argument
Individual observations rarely establish viral origin.
The more consequential scientific argument concerns their joint occurrence.
SARS-CoV-2 appeared with:
- a distinctive polybasic furin-cleavage site;
- a 12-nucleotide insertion at the S1/S2 boundary;
- the PRRA sequence;
- the CGG-CGG arginine codons within that insertion;
- an RBD capable of efficient interaction with human ACE2;
- a coronavirus genome related to viruses actively investigated before 2020;
- a geographic emergence in the same city containing a major bat-coronavirus research program; and
- no identified immediate progenitor that completely closes the evolutionary gap.
The appropriate statistical question is therefore not whether each individual observation can occur naturally.
It is whether the joint probability and explanatory coherence of the complete constellation are better explained by one origin model than another.
That question remains incompletely resolved.
11. What Evidence Would Resolve the Laboratory Hypothesis?
The laboratory hypothesis is unusually amenable to falsification because laboratories generate records.
A decisive investigation should seek:
Viral inventories. Every sarbecovirus specimen physically possessed before December 2019.
Raw sequencing data. Not merely published consensus sequences, but original reads from field specimens and laboratory isolates.
Passage histories. Documentation of every experimental propagation and adaptation experiment.
Reverse-genetics records. Plasmids, infectious clones, sequence files and experimental protocols.
Animal-study records. Species, inoculum, passage number, viral sequences and clinical observations.
Laboratory notebooks. Particularly records concerning unpublished sarbecoviruses.
Biosafety records. Occupational exposures, accidents, illnesses and serological testing among personnel.
Historical databases. Archived sequence repositories and laboratory databases capable of identifying previously unpublished viral material.
Environmental samples. Archived samples from relevant research facilities and surrounding environments.
These data would allow investigators to move beyond inference toward reconstruction.
12. A New Standard for Origins Research
The scientific debate should ultimately be reframed.
The objective should not be to demonstrate that one hypothesis is ideologically acceptable and another unacceptable. Nor should the objective be to transform incomplete evidence into certainty.
The objective should be forensic reconstruction.
Modern genomics permits extraordinarily detailed reconstruction of viral ancestry. Modern laboratory records permit reconstruction of experimental histories. Modern epidemiology permits reconstruction of transmission chains. Modern computational biology permits comparison of alternative evolutionary trajectories.
These capabilities should be combined.
The question is sufficiently important that the investigation should be treated with the same methodological seriousness applied to the reconstruction of a major industrial accident, an unexplained aircraft failure or an unusual outbreak of a high-consequence pathogen.
Conclusions
The laboratory-associated hypothesis for SARS-CoV-2 does not depend upon a single extraordinary observation.
Its scientific importance arises from the convergence of several observations: the unusual polybasic furin-cleavage site; the 12-nucleotide insertion that generates it; the nucleotide composition of that insertion; the biological importance of spike cleavage; the functional ACE2-binding RBD; the known pre-pandemic experimental capacity to manipulate and study sarbecoviruses; the existence of extensive coronavirus research in Wuhan; the possibility of laboratory passage or recombination; and the continuing absence of a fully resolved immediate progenitor.
None of these observations should be treated as a substitute for direct evidence.
But neither should the absence of a documented laboratory construct be treated as proof that no laboratory-associated event occurred.
The central scientific issue is provenance.
Where did the virus—or its immediate precursor—exist before it entered the human population?
If the answer is a wildlife population, the relevant evolutionary pathway should be reconstructable from the ecology and genetics of that reservoir. If the answer involves laboratory research, the relevant pathway should be reconstructable from viral inventories, experimental records, sequencing data and laboratory histories.
The most consequential unresolved question is therefore not whether SARS-CoV-2 “looks man-made.”
It is whether the complete molecular, epidemiological and laboratory record can establish how the virus acquired its defining biological characteristics and where that process occurred.
That is a question worthy of the highest standards of contemporary molecular epidemiology.
Footnotes and References
1. Andersen KG, Rambaut A, Lipkin WI, Holmes EC, Garry RF. The proximal origin of SARS-CoV-2. Nature Medicine. 2020;26:450–452. The paper identified the human ACE2-binding RBD and the polybasic furin-cleavage site as two notable features of SARS-CoV-2.
2. The SARS-CoV-2 furin-cleavage site consists of a polybasic RRAR sequence at the S1/S2 junction and results from a 12-nucleotide insertion. The biological location of this insertion makes it particularly relevant to investigations of viral entry and host adaptation.
3. The distinction between laboratory-associated emergence and intentional genetic engineering is fundamental. A laboratory-associated event could involve naturally occurring viruses, laboratory passage, recombination, adaptation or other manipulation without requiring construction of an entirely synthetic genome.
4. Prior to the COVID-19 pandemic, coronavirus research included investigation of bat SARS-like coronaviruses and mechanisms of cross-species transmission. The existence of this experimental capability establishes technological plausibility but does not by itself establish provenance. The scientific literature from the period documents the use of coronavirus reverse-genetics approaches.
5. The earliest genomic literature placed SARS-CoV-2 within the sarbecovirus lineage and noted substantial similarity to a bat coronavirus identified in 2013. The same literature established that recombination is an important feature of coronavirus evolution.
6. The WHO Scientific Advisory Group for the Origins of Novel Pathogens (SAGO) published an independent assessment in June 2025 after reviewing published literature, reports, scientific presentations, interviews and other available evidence. WHO stated that substantial information required to fully evaluate all hypotheses had not been provided and called for additional information concerning early cases, animal data and Wuhan laboratory activities.
7. WHO’s 2025 global framework explicitly recognizes investigation of laboratory biosafety and biosecurity as part of systematic pathogen-origin investigations.
8. A rigorous origins investigation should distinguish evidence demonstrating that a laboratory could have generated or acquired a virus from evidence demonstrating that a laboratory did generate or acquire that virus. This distinction is essential to maintaining scientific credibility.
9. The strongest laboratory hypothesis is therefore not necessarily “SARS-CoV-2 was designed as a weapon.” A broader laboratory-associated hypothesis encompasses accidental release following collection, culture, passage, adaptation, recombination or experimental manipulation of a naturally occurring sarbecovirus.
10. The ultimate evidentiary standard should be reconstruction: identification of a precursor virus, laboratory sequence, experimental record, passage history, or other primary evidence that establishes a continuous chain connecting a pre-pandemic virus to SARS-CoV-2.