For therapeutic proteins you usually work the other way around: You design the shape and try to get a sequence that folds as intended. In other words, the inverse folding problem.
While this is true, DeepFold’s algorithm is only applicable to extant proteins having enough evolutionary information, which is arguably a small fraction of the theoretical sequence space. Fortunately, machine-learning approaches for de novo protein design are being actively developed as we speak.
Didn't the challenge include predicting previously unseen/unsolved proteins? Based on that, I would wager that what DeepFold learned is an evolutionary "language" that maps out a large useful subset of the entire possibility space. Natural evolution tends build upon previous successes so it seems probable evolution has mapped out a fairly useful "language" of patterns for useful protein shapes. Especially given only a portion of the shape is critical to function for many proteins.
But agreed, de novo protein design based on DeepFold's successes probably will outperform a naive approach by orders of magnitude. But why wait if even a naive approach is already orders of magnitude better than current methods?
Either way, I'm excited to see who comes up with the first custom protein(s) to catalyze industrial processes! Get some yeast/bacteria to mass produce a protein based alternative to platinum catalyzer's for fuel cells using an active site with organically available (and cheap) metals. Design half of it to stick to a polymer so it coats nicely. Bam, no more trying to get some weird polymer/perovskite with the right properties. Not sure mRNA is needed at that point versus CRISPR, but maybe it's more effective.
All protein sequences in the competition lacked a published solved structure, but they had enough effective (remotely) homologous sequences to predict coevolution-derived interresidue distances and contacts. All those sequences were already present in databases and therefore within the known protein universe.
> Based on that, I would wager that what DeepFold learned is an evolutionary "language" that maps out a large useful subset of the entire possibility space.
This might be true, depending on how “foldable” the unexplored space is. A reverse DeepFold would give some clues on that.
> Natural evolution tends build upon previous successes so it seems probable evolution has mapped out a fairly useful "language" of patterns for useful protein shapes
This is already known. The structural and functional diversity we see in existing proteins originated from a relatively limited repertoire of conserved protein domain folds, or even subdomain-sized fragments in some cases.
> Either way, I'm excited to see who comes up with the first custom protein(s) to catalyze industrial processes!
Same here! Especially since the de novo design of enzymes has been progressing slowly but steadily.
Ah, I see what you're saying a bit better. Makes more sense to how the the possible "design" space for novel proteins could be limited. Designing novel active sites could be especially tricky, more than it'd seem at first glance. A folded structure that'd effectively transfer electrons from one target species to another in a catalyzer protein could likely be outside the explored "vocabulary" of extant proteins as it'd require specialized pathways and precise positioning. Chlorophyll is pretty unchanged in evolution as I understand it.Thanks, interesting background! I'll keep an eye out for the de novo design algorithms.
That said, I can imagine once we figure out how to deliver the mRNA to precisely the right targets, this would be a lot easier (eg engineering receptor ligands onto the mRNA lipid delivery particles).
I still have questions about the mRNA vaccine concept. Maybe some well-informed HNer can shed light?
1. If the goal is to get the desired protein inside the patient's body, why is injecting mRNA that produces the protein more effective than simply introducing the protein itself?
2. How does the patient's immune system know to treat this protein as unwanted if it's being produced by the body's own cells?
3. How do vaccine developers avoid creating an over-active immune response in the patient (i.e. an allergy)?
So the trick really is to design the RNA and the delivery system so that it produces an immune response of the right magnitude. Too weak and it doesn't work, too strong and it's harmful.
Antibodies that bind to your own proteins are filtered in an earlier stage in antibody production. Real viral proteins are also produced by your own cells, so that is never a factor by which you could distinguish friend from foe.
2. The imune system knows all your protein already, t cells are evolved in the thymus gland, if they attack anything "self", they get destriyed.
Which one? I don't remember any vaccine that injects immortalized cells.
the cells are used to manufacture quantities of the virus in an extremely weakened form. This gives a chance for the immune systemto mount a response before the pathogen initiates systemic disease response.
the cells in question are not freshly harvested from a foetus, the progenitor cells are harvested and kept in continual culture for use. the reason immortalized cells are used is to escape the "expiry date" of a normal cell.
>>Some vaccines currently available were developed using cell strains cultured from two fetuses aborted for other purposes in the 1960s.[6]<<
i lifted this link from the wiki page provided:
So far, nobody has been able to provide a satisfying answer.
The other mechanism is that the immune system doesn't react only to foreign antigens, there also needs to be some general activation of the immune system in that area for a real response.
The full answer is that this is really, really complex like pretty much everything involving the immune system. And auto-immune diseases indicate that the immune system doesn't always get this right.
in very loose terms both keys must be presented while the immune system is in training, the instances where protiens are expressed with out cell surface recognition molecules in proximity are invalid thus foriegn, and memory immunity to this "failed login" is maintained. when both factors [protien plus recognition] are present constituitively over the course of development the immune response is squelched.
this is known as central tolerance.
The immune system can't know who produced a certain molecule.
2. Someone already said, but a protein has no marker saying "local produce". It is either allowed or not and the body reacts or not when detecting it.
3. They hope. Measure the dosage and adjust. As it was reported, some people getting the vaccine have an allergic reaction, but hopefully it will be rare and under control. So far there are only a few cases reported.
My intuition is that many cancers can start when proteins are mis-copied, and mRNA therapy seems to rely on copying to work very reliably. Does introduced mRNA increase the risk or impacts of mis-copying?
Is there an ELI5 intuition for why mRNA therapy is safe and won't lead to cells becoming cancerous?
Same as always: do a long-term clinical trial and look for any side effects.
The ELI5 version:
Think of the cell as a factory for proteins. The blueprints for all of the different kinds of proteins it can produce are encoded in DNA and stored in the foreman's office. When the cell wants to make a particular protein, the foreman copies the instructions from the blueprint onto slips of paper (mRNA), and sends them to the assembly line workers. This type of copying happens all the time, and is pretty reliable -- and even if it's not, the end result is just some wasted time and a junk part. The blueprints are unaffected. mRNA vaccines sneak some unauthorized instructions to the workers, who then run a ghost shift to produce the protein we want to cause an immune response to.
When we need to create a second factory, we need to photocopy all the blueprints for the new cell. This uses a totally different process (a photocopier assembled specifically for the job of copying whole blueprints onto more blueprint paper, and not the work-order slips sent to the assembly line). The photocopier is pretty reliable too, but occasionally it introduces artifacts, and eventually, enough of these errors can build up to create mistakes in the blueprints. The blueprints have instructions on how to deal with errors in a new factory (by blowing it up), but if enough of those failsafe instructions are obscured by copier artifacts, you now have a potentially cancerous cell.
The foreman's office doesn't usually make changes to the blueprints based on work order slips. It's possible with a special copy machine -- this is how HIV works, by sneaking that machine in -- but without a work order slip to blueprint copier, it doesn't matter how many fake work order slips we pass into the cell. They will only affect the production line, not the foreman's office.
The more detailed, and almost certainly wrong-in-the-details version:
There are three types of encoding for genetic material: in DNA, in RNA, and in proteins.
In order for the information encoded in genetic information to be active in a biological system, it has to be expressed as a protein -- a physical manifestation with binding sites and the ability to interact with other molecules. Ribosomes, a component of the cell, do the "assembly" part of transferring genetic information to a protein. This is called "translation".
DNA is the long-term storage of the information eventually expressed as proteins -- it's a collection of templates in long term storage. So there has to be a process to get information from the long-term storage of DNA, in the nucleus of the cell, to the ribosomes elsewhere in the cell.
Messenger RNA is the intermediary molecule which carries the genetic information from the DNA to the ribosomes. It's produced in the nucleus of the cell by a mechanism called "transcription", which copies a subset of a DNA molecule into a complementary RNA molecule encoding the same information. The DNA double-helix is "unzipped", a complementary RNA strand is assembled, and the double-helix reforms. The mRNA can then be transported to the ribosomes.
mRNA is a useful vaccine mechanism, because we're co-opting a late stage of the cell's protein production pipeline. We don't have to assemble the target protein by hand and figure out how to modify it make it stable for distribution -- a task that would need to be repeated for each protein we wanted to make. Instead, we can figure out how to synthesis mRNA, and we don't need to worry about either synthesis or stability for the end protein product.
When we talk about cancer and mis-copying, this is at a much earlier stage in the pipeline. When a cell splits into two cells through mitosis, we don't want a subset of the protein templates in the cell's genetic material; we want a complete copy. This is a separate biological pipeline, which involves the cell building up a mechanism to do the copying, dissolving the membrane of the nucleus, and ultimately producing and collating copies of all the DNA from the nucleus. This is an entirely separate process, and messenger RNA isn't involved at all.
There is a route for information to travel from RNA back to DNA, and this is the mechanism retroviruses use, and HIV is the most famous example. A complete retrovirus particle includes reverse transcriptase, an enzyme which allows them to transfer information from RNA back to DNA. This is the reverse of the normal DNA -> RNA pattern described above, hence the name _retro_virus. Without that enzyme or something like it, there isn't a pathway to convert RNA back to DNA, and human cells don't normally express an enzyme which has this function.