3 ms·

Impressive work, but author's logic is unsound because it silently assumes the library or database BLAST uses is complete, yet it likely represents less than 1%

Impressive work, but author's logic is unsound because it silently assumes the library or database BLAST uses is complete, yet it likely represents less than 1% of all sequences in the wild. The database BLAST uses only contains sequences that have been found and reported. If every person on the planet was a genetic expert with all necessary resources to find and identify new sequences, in 1000 years we still wouldn't have found all of them. The sequence in question that only matches HIV-1 may very well match a trillion^2 more sequences not in BLAST's database. Anything like this that is engineered likely has a mountain of documentation. Documentary evidence, if found, is a smoking gun in a way the furin cleavage site sequence can't be.

WanderPanda 5y ago
It would be interesting to me how likely it is to actually find random sequences in the database. Aka how much correlation there is between natural viruses. In my imagination it is plausible that if you take a particular virus that you will be able to find most of its subsequences (with minor modifications) in other viruses. In that case I think that his analysis is sound. On the other hand, if the other viruses appear to be randomly sampled from the set of all possible sequences than his reasoning does not cut it.
cuteboy19 5y ago
Is there any reason to believe that the database isn't a random sample of all possible sequences?
- etaoins 5y ago
  Genomes are sequenced when they're both accessible and are of scientific/economic interest. This results in an overrepresentation of: - Humans and their pathogens - Economically important organisms (e.g. crops and farm animals) and their pathogens - Model organisms For example, compare the number of Genebank sequences for HIV-1 [1] versus Feline immunodeficiency virus [2]. HIV-1 is additionally problematic because it has an extremely high mutation rate [3], making it more likely for a random sequence to match some sequenced HIV-1 genome. [1]: https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Tree&id=11676&lvl=3&lin=f&keep=1&srchmode=1&unlock https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mod... [2]: https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Tree&id=11673&lvl=3&lin=f&keep=1&srchmode=1&unlock https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mod... [3]: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4574155/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4574155/
- fastaguy88 5y ago
  The HIV virus has been sequenced 10's of thousands of times, and all those sequences are in the database that was searched. Likewise, there are thousands of copies of the E. coli proteome in the sequence databases. The nr database, part of which was searched, is extremely redundant (despite the fact that 'nr' stands for "non-redundant" -- no two identical sequences). So unless you go to a lot of trouble (e.g. the NCBI landmark database), most sequence databases are a mixture of 10's to 10,000's of very closely related sequences (not a random sample), from a non-random sample of organisms (we tend to sequence pathogens). Protein and DNA sequence databases are very large, but they are not random.

WanderPanda 5y ago

It would be interesting to me how likely it is to actually find random sequences in the database. Aka how much correlation there is between natural viruses. In my imagination it is plausible that if you take a particular virus that you will be able to find most of its subsequences (with minor modifications) in other viruses. In that case I think that his analysis is sound. On the other hand, if the other viruses appear to be randomly sampled from the set of all possible sequences than his reasoning does not cut it.

cuteboy19 5y ago

Is there any reason to believe that the database isn't a random sample of all possible sequences?

etaoins 5y ago
Genomes are sequenced when they're both accessible and are of scientific/economic interest. This results in an overrepresentation of: - Humans and their pathogens - Economically important organisms (e.g. crops and farm animals) and their pathogens - Model organisms For example, compare the number of Genebank sequences for HIV-1 [1] versus Feline immunodeficiency virus [2]. HIV-1 is additionally problematic because it has an extremely high mutation rate [3], making it more likely for a random sequence to match some sequenced HIV-1 genome. [1]: https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Tree&id=11676&lvl=3&lin=f&keep=1&srchmode=1&unlock https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mod... [2]: https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Tree&id=11673&lvl=3&lin=f&keep=1&srchmode=1&unlock https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mod... [3]: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4574155/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4574155/
fastaguy88 5y ago
The HIV virus has been sequenced 10's of thousands of times, and all those sequences are in the database that was searched. Likewise, there are thousands of copies of the E. coli proteome in the sequence databases. The nr database, part of which was searched, is extremely redundant (despite the fact that 'nr' stands for "non-redundant" -- no two identical sequences). So unless you go to a lot of trouble (e.g. the NCBI landmark database), most sequence databases are a mixture of 10's to 10,000's of very closely related sequences (not a random sample), from a non-random sample of organisms (we tend to sequence pathogens). Protein and DNA sequence databases are very large, but they are not random.