[math-fun] Spelling correction (ASLR?) on DNA sequences
It appears that using "CRISPR" technology, it is possible to "correct the spelling" of various DNA *protein coding subsequences* to replace all coding triplets for a particular amino acid with a single, canonical triplet for that amino acid. This "spelling correction" of an organisms's genome apparently *does not harm* the creature, since precisely the same proteins are still constructed in precisely the same way. However, this "spelling correction" seems to play havoc with a majority of the usual viruses that want to attack the organism, and may provide some degree of immunity -- at least until the viruses evolve to recognize & take advantage of this new/canonical coding triplet sequence. This "spell checking" is analogous to, but simpler than, the "Address Space Location Randomization" which is used to stymie computer computer viruses & malware. Of course, ASLR-type randomization may soon be required in humans & other organisms when the biological viruses begin to recognize a canonical coding. Speaking of viruses, it also appears that tens of *complete viruses* are encoded in animals' DNA -- e.g., pigs -- in such a way that theses viruses can escape from the DNA and become fully active. Typically, these pigs are already immune (or partially immune) to these hidden viruses, but *other organisms* -- e.g., humans -- who come in contact with them may not be. This "hidden virus" problem makes transplanting *pig organs* into humans much more difficult, as transplantation requires *immunosuppression*, which then makes the human recipient completely vulnerable to these hidden viruses. Apparently, CRISPR can also be used to remove these hidden viruses, and thus produce pigs whose organs don't harbor such viruses. https://after-on.com/episodes/024 Episode 24: George Church | Bioengineering After On Podcast #24: The Astounding Present and Dizzying Future of Synthetic Biology (from Boing Boing) George Church's Harvard lab is one of the most celebrated fonts of innovation in the world of life sciences. George's earliest work on the Human Genome Project arguably pre-dated the actual start of that project. Subsequently, he's been involved in the creation of almost a hundred companies - 22 of which he co-founded. Much of George's most recent and celebrated work has been with a transformationally powerful gene-editing technique called CRISPR, which he co-invented. ...
I believe there are several problems with this discussion. I’ll try to elaborate.
On May 6, 2018, at 11:34 PM, Henry Baker <hbaker1@pipeline.com> wrote:
It appears that using "CRISPR" technology, it is possible to "correct the spelling" of various DNA *protein coding subsequences* to replace all coding triplets for a particular amino acid with a single, canonical triplet for that amino acid.
This "spelling correction" of an organisms's genome apparently *does not harm* the creature, since precisely the same proteins are still constructed in precisely the same way.
This is wildly optimistic. There are many issues in the choice of DNA sequence having little to do with the protein coding, such as folding properties, accidental encoding of other DNA sequences, and overuse of cognate tRNAs for a particular codon. Almost certainly this could work for a single protein, with some chance, but would almost certainly fail if tried on a large scale.
However, this "spelling correction" seems to play havoc with a majority of the usual viruses that want to attack the organism, and may provide some degree of immunity -- at least until the viruses evolve to recognize & take advantage of this new/canonical coding triplet sequence.
Viruses don’t depend on the coding of cell proteins for their activity. They depend on being able to co-opt the cell mechanisms to make their own proteins. You may have been fooled by the wild claims of viral immunity based on the idea of simply removing the tRNAs for specific codons entirely from the genomes of organisms. This has been done by Church in E. coli, but there is a huge difference between doing this in a 4e6 genome organism (and a relatively simple one at that) and a 3e9 genome such as you. Not that we have a clue as to how to do that massive a change in you.
This "spell checking" is analogous to, but simpler than, the "Address Space Location Randomization" which is used to stymie computer computer viruses & malware. Of course, ASLR-type randomization may soon be required in humans & other organisms when the biological viruses begin to recognize a canonical coding.
Speaking of viruses, it also appears that tens of *complete viruses* are encoded in animals' DNA -- e.g., pigs -- in such a way that theses viruses can escape from the DNA and become fully active. Typically, these pigs are already immune (or partially immune) to these hidden viruses, but *other organisms* -- e.g., humans -- who come in contact with them may not be.
This "hidden virus" problem makes transplanting *pig organs* into humans much more difficult, as transplantation requires *immunosuppression*, which then makes the human recipient completely vulnerable to these hidden viruses.
Apparently, CRISPR can also be used to remove these hidden viruses, and thus produce pigs whose organs don't harbor such viruses.
Well, yes, but this is only a very small part of what will be necessary.
https://after-on.com/episodes/024
Episode 24: George Church | Bioengineering
After On Podcast #24: The Astounding Present and Dizzying Future of Synthetic Biology (from Boing Boing)
George Church's Harvard lab is one of the most celebrated fonts of innovation in the world of life sciences. George's earliest work on the Human Genome Project arguably pre-dated the actual start of that project. Subsequently, he's been involved in the creation of almost a hundred companies - 22 of which he co-founded. Much of George's most recent and celebrated work has been with a transformationally powerful gene-editing technique called CRISPR, which he co-invented.
You should have a studied skepticism of everything George says. Much of the work is very good, but it is typically placed in a wildly optimistic and expansive context, often misleading.
...
_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
BTW, these ideas weren't mine; I was merely reporting them from George Church's discussion in the podcast link below. Only the "ASLR" analogy is mine. If I understood George correctly, only the triplets *actually coding for an amino acid* would have their spelling "corrected", and this % of all DNA is quite small -- perhaps < 10% ? Thus the other "some assembly required" instructions DNA wouldn't be affected by these changes. The ASLR analogy is similar: ASLR only helps a little against malware, and works only because a lot of computer code doesn't care *where* the storage is, only that it exists and that addressing within a chunk works. Obviously, randomizing executable code itself -- e.g., to combat "gadgets" and "ROP" (Return-Oriented Programming) -- would require a good deal more subtlety -- e.g., "PIC": "position independent code", etc. At 05:50 AM 5/7/2018, Tom Knight wrote:
I believe there are several problems with this discussion.
I'll try to elaborate.
On May 6, 2018, at 11:34 PM, Henry Baker <hbaker1@pipeline.com> wrote: It appears that using "CRISPR" technology, it is possible to "correct the spelling" of various DNA *protein coding subsequences* to replace all coding triplets for a particular amino acid with a single, canonical triplet for that amino acid.
This "spelling correction" of an organisms's genome apparently *does not harm* the creature, since precisely the same proteins are still constructed in precisely the same way.
This is wildly optimistic. There are many issues in the choice of DNA sequence having little to do with the protein coding, such as folding properties, accidental encoding of other DNA sequences, and overuse of cognate tRNAs for a particular codon. Almost certainly this could work for a single protein, with some chance, but would almost certainly fail if tried on a large scale.
However, this "spelling correction" seems to play havoc with a majority of the usual viruses that want to attack the organism, and may provide some degree of immunity -- at least until the viruses evolve to recognize & take advantage of this new/canonical coding triplet sequence.
Viruses don't depend on the coding of cell proteins for their activity. They depend on being able to co-opt the cell mechanisms to make their own proteins. You may have been fooled by the wild claims of viral immunity based on the idea of simply removing the tRNAs for specific codons entirely from the genomes of organisms. This has been done by Church in E. coli, but there is a huge difference between doing this in a 4e6 genome organism (and a relatively simple one at that) and a 3e9 genome such as you. Not that we have a clue as to how to do that massive a change in you.
This "spell checking" is analogous to, but simpler than, the "Address Space Location Randomization" which is used to stymie computer computer viruses & malware. Of course, ASLR-type randomization may soon be required in humans & other organisms when the biological viruses begin to recognize a canonical coding.
Speaking of viruses, it also appears that tens of *complete viruses* are encoded in animals' DNA -- e.g., pigs -- in such a way that theses viruses can escape from the DNA and become fully active. Typically, these pigs are already immune (or partially immune) to these hidden viruses, but *other organisms* -- e.g., humans -- who come in contact with them may not be.
This "hidden virus" problem makes transplanting *pig organs* into humans much more difficult, as transplantation requires *immunosuppression*, which then makes the human recipient completely vulnerable to these hidden viruses.
Apparently, CRISPR can also be used to remove these hidden viruses, and thus produce pigs whose organs don't harbor such viruses.
Well, yes, but this is only a very small part of what will be necessary.
https://after-on.com/episodes/024
Episode 24: George Church | Bioengineering
After On Podcast #24: The Astounding Present and Dizzying Future of Synthetic Biology (from Boing Boing)
George Church's Harvard lab is one of the most celebrated fonts of innovation in the world of life sciences. George's earliest work on the Human Genome Project arguably pre-dated the actual start of that project. Subsequently, he's been involved in the creation of almost a hundred companies - 22 of which he co-founded. Much of George's most recent and celebrated work has been with a transformationally powerful gene-editing technique called CRISPR, which he co-invented.
You should have a studied skepticism of everything George says. Much of the work is very good, but it is typically placed in a wildly optimistic and expansive context, often misleading.
...
participants (2)
-
Henry Baker -
Tom Knight