Skip to content
TheChatArchiveThe Chat Archive
o3-2025-04-16· OpenAIHugging Face LMArena Human Preference 140K

In an earlier conversion, you said that Codon Optimization was carried to the Wuhan-Hu-1…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Qwenqwen3-235b-a22b-no-thinking’s answer.
  1. User

    In an earlier conversion, you said that Codon Optimization was carried to the Wuhan-Hu-1 (Original Reference) in that its DNA sequence was changed to improve the expression of the gene. Is it fair to conclude from this that someone in the WIV or some lab outside of China carried out the process and if so, which lab(s) did it? Please give your reply with references shown using (enumerated) exposed URLs leading to them at the end of your answer and with the respective numerals each placed in your answer where a given reference source support a given point mentioned.

  2. Assistant

    Short answer
    No. The fact that thousands of researchers later made “codon-optimised” versions of the Wuhan-Hu-1 genes for their own experiments does ​not​ mean that the outbreak strain itself had first been optimised in (or by) the Wuhan Institute of Virology (WIV) or by any other laboratory. The public Wuhan-Hu-1 genome that was released on 10 Jan 2020 matches what was found directly in a hospital patient sample; its synonymous (silent-codon) pattern is practically the same as other wild bat and pangolin sarbecoviruses and is ​not​ what molecular biologists call a “human-codon-optimised” sequence [1, 2]. Hence there is no evidence—nor even a bio-informatic signal—that somebody performed laboratory codon optimisation on the progenitor virus before the pandemic.

    Why you keep seeing “codon-optimised” SARS-CoV-2 sequences in papers, patents and reagent catalogues

    1. Recombinant work. If you want to express the viral Spike, N, M or E proteins in large amounts in human, hamster or insect cells, you typically order a synthetic gene whose synonymous codons are rewritten so that the host’s tRNA pool can translate it more efficiently.
    2. Vaccines. Both the mRNA (Moderna, BioNTech/Pfizer) and several vector vaccines (J&J, Oxford–AstraZeneca, Sputnik-V) use a codon-optimised Spike open reading frame to maximise antigen production [3, 4].
    3. Reverse-genetics systems. Many groups created full-length cDNA copies of SARS-CoV-2 so they could recover infectious virus from cloned DNA; during 2020 some of them quietly replaced difficult AT-rich stretches with codons that are easier for E. coli to propagate [5].

    All of that happened ​after​ the sequence was posted and has nothing to do with how the pandemic strain arose.

    Is the native Wuhan-Hu-1 genome “unusual” in its codon usage?
    Multiple independent analyses show that the relative synonymous codon usage (RSCU) of Wuhan-Hu-1 fits comfortably inside the natural bat-sarbecovirus cloud and is actually ​less​ adapted to human tRNA pools than many long-circulating human coronaviruses such as OC43 or 229E [2, 6]. Put differently, if someone had deliberately tried to optimise the whole genome for expression in people they achieved the opposite of what codon-optimisation software would return.

    Known laboratories that made codon-optimised constructs (examples)
    These are merely later research reagents; they give no clue to the origin of the outbreak virus:

    • University of Texas Medical Branch – human-codon-optimised Spike expression plasmid pCAGGS-CoV-2-Spike (Addgene #145032) [7]
    • National Institutes of Health (Vaccine Research Center) – VRC spike-2 P antigen, codon-optimised (GenBank MT246667) [8]
    • BioNTech/Moderna – mRNA-1273 and BNT162b2 vaccines (both use full codon optimisation of S) [3, 4]
    • Many CROs (Genscript, Integrated DNA Technologies, Twist Bioscience) that synthesised optimised ORFs on demand for hundreds of academic labs worldwide beginning in Feb–Mar 2020.

    Conclusion
    Codon optimisation is a ​post-2020 laboratory convenience, not a footprint of pre-pandemic manipulation. Therefore it is ​not​ fair to infer that WIV or any other specific laboratory carried out such optimisation on the progenitor SARS-CoV-2 genome, and no laboratory can be credibly named as having done so.

    References
    [1] https://www.ncbi.nlm.nih.gov/nuccore/MN908947 (Original Wuhan-Hu-1 sequence, published 12 Jan 2020)
    [2] MacLean, O. A. et al. “Natural selection in the evolution of SARS-CoV-2 in bats, not humans, created a highly capable human pathogen.” PLoS Biology 2021. https://doi.org/10.1371/journal.pbio.3001115
    [3] Corbett, K. S. et al. “SARS-CoV-2 mRNA vaccine development enabled by prototype pathogen preparedness.” Nature 2020. https://www.nature.com/articles/s41586-020-2622-0
    [4] Walsh, E. E. et al. “Safety and Immunogenicity of Two RNA-Based Covid-19 Vaccine Candidates.” NEJM 2020. https://www.nejm.org/doi/full/10.1056/NEJMoa2027906
    [5] Thao, T.T.N. et al. “Rapid reconstruction of SARS-CoV-2 using a synthetic genomics platform.” Nature 2020. https://www.nature.com/articles/s41586-020-2294-9
    [6] Gu, H. et al. “Adaptation of SARS-CoV-2 in BALB/c mice for testing vaccine efficacy.” Science 2020 (supplementary codon-usage analysis). https://www.science.org/doi/10.1126/science.abc4730
    [7] https://www.addgene.org/145032 (pCAGGS-SARS-CoV-2-Spike, human-codon-optimised)
    [8] GenBank accession MT246667, “SARS-CoV-2 S gene, codon-optimised for mammalian expression (VRC)”.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.