Comment by 110bpm | Hacker News Reader

Comment by 110bpm | Hacker News Reader

Parent Full thread

I wanted to learn more about where this claim comes from and found this article [0].

The linked commentary [1] raises several interesting points.

"Unfortunately, the authors did not provide details of their BLAST research. The patent database that they used contained 24,712 sequences. Yet, by querying BLAST, we obtained a database of 46,121,617 patent sequences with an average length of 560 nucleotides. The authors should give more details and justification for their query, especially if they queried the full database but a posteriori restricted their computation. Of note, with such a large database, and despite the fact that the average sequence length decreased, the probability of finding at least one sequence containing one of the 16 patterns previously mentioned may rise to 68% under assumption 2."

0: https://www.frontiersin.org/journals/virology/articles/10.33...

1: https://www.frontiersin.org/journals/virology/articles/10.33...

FTA:

The absence of CTCCTCGGCGGGCACGTAG from any mammalian or viral genome in the BLAST database makes recombination in an intermediate host an unlikely explanation for its presence in SARS-CoV-2. You can verify for yourself that there are no other sequences with an exact match.

The subsequent comment about statistics of blast searching is irrelevant.

So the conclusion of the comment-article cited in the parent of your comment is false?

"According to the current phylogeny, FCS appeared independently six times in the Betacoronavirus lineages, demonstrating that FCS insertion is compatible with natural evolution (2, 7, 8). [...] 7. Wu Y, Zhao S. Furin cleavage sites naturally occur in coronaviruses. Stem Cell Res (2021) 50:102115. doi: 10.1016/j.scr.2020.102115"

it's incompatible. Why is there COVID speculation in Stem Cell Res?

"Unfortunately, the authors did not provide details of their BLAST research. The patent database that they used contained 24,712 sequences. Yet, by querying BLAST, we obtained a database of 46,121,617 patent sequences with an average length of 560 nucleotides."

is that not relevant?

just do the BLAST yourself and see. it's a publically available database. you will find zero other exact matches. note that today in 2025 there are a shit tonne of sequences that are resequencing of covid variants so you'll have to filter those out

https://blast.ncbi.nlm.nih.gov/Blast.cgi

Given we are on HN, it seems more likely that the claim originated on substack. https://news.ycombinator.com/item?id=29938732 this may be one of the progenitor threads.

Reply on news.ycombinator.com