Please log in.
Before you can vote, you need to register. Please log in or create an account.
Science: Health: Coronavirus
Viral Immunity by Leukocyte Exposure   (+1)  [vote for, against]
A (possibly) stupid idea of what do do when you don’t have a vaccine for a novel virus

OK... here goes. This might be a deeply flawed idea but it’s been rattling around my brain-case for a while so I thought I’d send it out into the world.

**Situation**
A novel virus exists for which we don’t have a vaccine.

**Assumptions**
-1- A person who has been exposed to, become infected by, and who has recovered from a virus will have future resistance to that virus.
-2- The mechanism by which we develop and express that resistance will be driven by leukocytes.
-3- Leukocytes are for all intents and purposes immune to viral infection because their purpose is to learn from and fight antigens.
-4- Once a leukocyte bound to a virus, it will not let it go. (Not critical this one, but it would change the method if leukocytes shed intact virions that they have learned).

If we can expose some of a person’s leukocytes to a live virus without exposing the rest of their cells then person can gain resistance to the virus without becoming infected or contagious.

**Method**
-1- Remove some blood from the patient.
-2- Fractionate to separate leukocytes.
-3- Create conditions that will encourage leukocytes to bind to antigens.
-4- Expose leukocytes to virus for long enough for to allow leukocytes to attach to virions.
-5- Wash/fractionate leukocytes repeatedly to remove unattached virions.
-6- Mix leukocytes back with other blood fractions and reintroduce to the patient.

If the leukocytes continue hold the virus to which they have bound themselves and develop an antigen without releasing the virus then the person can develop an immunity to the virus without getting infected.

**Notes**
-1- I don’t see this as being as effective (or as easy to administer) as a vaccine so it would only be of use when there is no vaccine currently available.
-2- I have generalised all cells involved in gaining immunity as leukocytes. I know that there are many types and subtypes but have kept this general to...
-a- ...keep this simple.
-b- ...avoid exposing the world to the extent of my ignorance.
-3- Lots of hand-wavey science. Sorry about that.
-- st3f, Apr 21 2020

Leuk, I am your Plasma https://www.thelanc...20)30141-9/fulltext
follow the bouncing B-lymphocytes [Sgt Teacup, Apr 22 2020]

// their purpose is to learn from and fight antigens. // // Once a leukocyte bound to a virus, it will not let it go.//

Welll ... no.

Antigens bond to proteins expressed on the surface of the virus's glycoprotein "coat" (for the class of virus currently causing alarm and despondency). This cripples the virus; with its bonding sites blocked, it cannot invade host cells and inject its RNA so as to reproduce.

The other "end" of the antigen then acts as a "flag" to leucocytes, indicating "eat me". The leucocyte envelops the now-inerted virus and digests it.

This process also triggers the production of more of the same antigens by the bone marrow, via a complex series of intermediary or messenger compounds.

Soon, the body is awash with more antigens, seeking matching jigsaw-puzzle sites to snap on to. Any free virus particles get drenched in antigens, then destroyed.

When the virus mutates, there will always be some "close enough" antigens that will latch on, triggering a new wave of digestion and antigen production. But this explains the "re-infection" phenomenon.

There are several sorts of vaccines; the ones for viruses typically rely on supplying a dose of "typical" antigens which act as a pattern to trigger the immune system when activated. The body keeps them around as it recognises "This is an antigen, leave it alone". But in some cases the awareness can decay, and periodic boosters are needed.

//If the leukocytes continue hold the virus to which they have bound themselves and develop an antigen without releasing the virus then the person can develop an immunity to the virus without getting infected. //

Sorry, but it doesn't work like that.

We trust our explanation is sufficiently clear. It is far from the real explanation, but is pitched at a level that the average reader might have a chance of understanding.

// avoid exposing the world to the extent of my ignorance //

It's a little late for that ...
-- 8th of 7, Apr 21 2020


When you said 'Antigens bond to proteins expressed on the surface of the virus's glycoprotein "coat"', did you mean to say Antibodies? Because my feeble understanding leads me to believe that the antigen is part of the virus.

Pardon my feeble ignorance in the light of your so obviously world rendering mighty intelligence.
-- st3f, Apr 21 2020


Antibodies bond mostly to bacteria. There are intermediate entities such as bacteriophages, part-way between viruses and bacteria, that react to either.

They fall into the broad category of "antigen reaction". We did point out that it was a deliberately simplistic and incomplete explanation.

Perhaps {MB] would be good enough to .... oh.

<Stares mournfully into middle distance/>

<Plods away/>
-- 8th of 7, Apr 21 2020


[st3f], we're working on this idea, sort of (see link).

B-lymphocytes (a type of leukocyte, working from home out of the humerus) produce antibodies to Covid19, but not well, and for some reason only in small amounts. Plus, about every 4 weeks the virus mutates, due to range of hosts (travel, social crowding, no water for hand-washing in crowded conditions) so we have make sure the plasma is coming from someone who got the same strain of Covid as the victim, er, patient.
-- Sgt Teacup, Apr 22 2020


// about every 4 weeks //

Researchers at a major European pharma company have recently ( last few days) characterised the half-life for mutation at 18 days; >90% divergence after 4 - 5 weeks, but after just 3 weeks any serum test will return an unacceptable proportion of false negatives, and ditto for any vaccine.

The mutations appear to be largely idiopathic, arising from transcription errors in the host cells during replication. No host-to-host process seems to be needed, although as you point out, that doesn't help.

Are there any portions of our previous explanation that you consider require revision ?
-- 8th of 7, Apr 22 2020


Would a plasma cocktail (shaken, not stirred), of several of the antibodies produced against several of the variants, be of some use?
-- RayfordSteele, Apr 22 2020


//portions... explanation... require revision//: Nah, that's good.

Observing JohnQPublic's inability to understand basic* immunology, 'I despair'**.

*Basic, like "Germs are not 'invisible', TrumpftyDumpty, we have microscopes for that".

**The shortest sentence in the world's most dangerous fictional history book is 'Jesus wept'. Same thing.
-- Sgt Teacup, Apr 22 2020


One of the reasons we have no vaccine or immunotherapy against the common cold (also a coronavirus) is because it's a quick, cagey bugger.

The Covid19 has an Achilles' heel: the deadlierist pneumonia arm (slower to mutate relative to the common cold portion, but mutates faster in confined, moist spaces like nursing homes and prisons).

A cocktail (stirred, not shaken) would need to contain local strains, where local can be determined by contact tracing.

If you put all known strains in the mix, introducing them to each other, you risk the unintended consequences of recombination.

Looks like the future is a small world after all.
-- Sgt Teacup, Apr 22 2020


// cocktail .... produced against several of the variants, be of some use? //

Yes, but as pointed out, only while the variation is not overly large; and the problem is lead time.

All viruses diverge in all directions, randomly and simultaneously. By the time a vaccine blend has been mass-produced and distributed (which takes days at very best, weeks in reality) the target has changed again. You may confer some small degree of preparedness in the body but not the solid immumological response needed to protect a vulnerable patient.

The vaccine needs to be administered well before exposure. That's why you have your jabs weeks before you go overseas. Once you have the infection it's too late.

The vaccine needs to be a fairly close match to the infective agent.

The vaccine must not transfer live agent to the recipient. Ensuring this is the case is a complex, multi-step process. Influenza vaccine is created in and extracted from chicken embryos, and it's "non-trivial" <link>.

And what [Sgt_T] said.

// in the world's most dangerous fictional history book is 'Jesus wept'. //

We searched the Qran, couldn't find that anywhere ...
-- 8th of 7, Apr 22 2020


Does the human immune system mutate it's own antibodies/immunoglobulins and antigen receptors to preempt the natural change in the pathogen antigens?
-- wjt, Apr 23 2020


Preemption is not an appropriate term as it implies a level of reasoning. However, some variations arise spontaneously and these can form the basis of an effective immune reaction against a novel pathogen.

The downside is that occasional - comparatively rare- malfunctions in this valuable and necessary function can give rise to very dangerous and damaging autoimmune reactions, where the system misidentifies "self" as "enemy".

As usual this is a gross oversimplification of a staggeringly complex process.
-- 8th of 7, Apr 23 2020


I imagine kids have this spontaneity in spades.
-- wjt, Apr 23 2020


Yes, and animals- even domesticated ones.

Watch a video of a cow getting a caesarian section - in a filthy barn, surrounded by straw, dirt, other cows, and the vet often doesn't wear gloves, nor a mask. Abdomen cut open; calf pulled out; abdomen stitched up; a bit of antibiotic spray on the wound.

Result ? Cow rapidly recovers. Animals have incredibly robust immune systems compared to humans ... though pedigree dogs, and thoroughbred horses, show the same traits.

Companion animals are interesting. They develop diseases never seen in the wild population, because they live two, three, five times longer than their wild counterparts. Their diseases are largely those of very old age. It's notable that animals in zoos live often twice the lifespan of their wild counterparts, often much more.

A problem arises when extraordinary efforts "save" human infants that are not "100%". They grow to adults and reproduce, diluting the gene pool. In the long term, this weakens the species as a whole.

In nature, sickly offspring die quickly. Nature is very, very harsh on the unfit. Tinker at your peril.
-- 8th of 7, Apr 23 2020


This comment is dedicated to corrections of what I consider obvious errors in the comments above only. It may not be complete, and I may of course make my own mistakes.

re. antigens, st3f is correct.
An antigen is a molecular structure which elicits an immune response. The thing which recognises those are antibodies. 8/7 is however correct that the immune system is horrendously complicated.

//There are intermediate entities such as bacteriophages, part-way between viruses and bacteria, that react to either. //

Bacteriophages are viruses which attack bacteria, not intermediates. Clue's in the name.

//All viruses diverge in all directions, randomly and simultaneously.//

While this is of course true (as it is of all life), it's not always the critical issue in vaccine creation that you've repeatedly been making it out to be. Most mutations are deleterious, so mutants are less fit, and tend not to propagate well. This is not uniform over the genome. A vaccine therefore ideally targets a part of the disease organism which can't change very much because of biological constraints.

//The vaccine needs to be administered well before exposure. [...] Once you have the infection it's too late.//

Not true, at least in general. Regarding the rabies vaccine, for example, the wikipedia page says "Treatment after exposure can prevent the disease if given within 10 days. The rabies vaccine is 100% effective if given early, and still has a chance of success if delivery is delayed"
I do concede it's not ideal, especially for rapid-onset diseases.

//A problem arises when extraordinary efforts "save" human infants that are not "100%". They grow to adults and reproduce, diluting the gene pool. In the long term, this weakens the species as a whole.//

Risky claim. Less fit genetically, but 'rescued' is likely to remain less fit overall. Therefore no genetically encoded deleterious trait is likely to achieve fixation in a large population. I doubt holding on to a bit more diversity will become a big problem.
-- Loris, Apr 23 2020


// re. antigens, st3f is correct. //

Corrections noted; the confusion arose as a result of extracting information from a rather long and extremely technical document.

The original errors will be left as a horrible example to others.

The comment on bacteriophages arose from the same document, which discusses a novel therapeutic regime which addresses similar "rapid change" problems in a different context.

We did say "in the long term". It's only been possible to do such things in the last century or so; even in Victorian times, infant mortality - particularly in cities - was of staggering proportions.

The effect is calculable from the observed effects on other species. Dogs have a "generation" of 2 to 4 years, horses 3 to 7. Many "purebred" dogs, selectively bred to enhance "desirable" traits, exhibit increasingly severe health issues that are very difficult to address. The times for the "classic" races have been dropping steadily for decades as the thoroughbred population, a closed gene pool, becomes smaller and smaller. That's a very extreme example, however.

Large interventions, like in-utero surgery, can save a fetus that would otherwise have died, in anything other than a highly technologically advanced and wealthy society that can afford to spend the resource to do that. The question of whether it should be done remains unanswered.

One scenario is that as these interventions become possible, they also become more necessary to support "normal" reproduction. Fertility drops, more children need "help" to survive.

Then the need for pre-reproduction genetic screening, to thwart the development of high-risk embryos, starts to become needed. Wow, won't that be popular.

It is likely that, through the Law of Unintended Consequences and a perfectly understandable desire to "do good" for offspring, your species may slowly paint itself into a corner, genetically.

Please point out any flaws in this argument. We are interested to hear your opinion.

To continue:

// Most mutations are deleterious, so mutants are less fit, and tend not to propagate well. This is not uniform over the genome. //

This is entirely correct, and would be relevant if it were not for (a) the astronomic numbers of viruses released by an infected host, and (b) the very high transmissibility.

The genome is very small, and RNA based. Unlike DNA, there are no checks and balances in replication so the probability of a single bit error is high. 99.99% of those errors cripple the virus produced, and it goes no further. But since the numbers produced are so staggering, a few variations survive and thrive.

// A vaccine therefore ideally targets a part of the disease organism which can't change very much because of biological constraints. //

... except as [Sgt_T] points out, this constant shape-shifting is the exact mechanism the virus relies on. "Make many, many, poor quality copies, quite a few slightly different. Some will survive". And so they do.

It's a successful bug because it is highly infectious but actually kills very, very few of its hosts.

// Not true, at least in general. Regarding the rabies vaccine, for example, the wikipedia page says "Treatment after exposure can prevent the disease if given within 10 days. The rabies vaccine is 100% effective if given early, and still has a chance of success if delivery is delayed" //

There are two sorts of rabies vaccine. One is given pre-exposure, then an antibody titre is used after an interval to check that immunity has been established. The second is given post-exposure and treatment must be given as soon as possible to trigger antibody production before the incubation period has elapsed. Effectiveness is very good, but once symptoms develop the chances of survival are very poor (but improving). Rabies is a poor fit for the Covid model as it's usually very clear when a patient has been exposed, plus the incubation period is comparatively long.

// I do concede it's not ideal, especially for rapid-onset diseases. //

It's not the rapid onset, it's the covert infectivity, plus 50%+ of cases are infective but entirely asymptomatic.

It appears that an effective treatment, if used early enough, is to give the patient a transfusion of plasma from a patient who has had the virus and recovered. However, the plasma must have a very low or nil load of live virus, and the strains in both donor and recipient must be a really "good" match. Transfusing plasma from an random survivor/asymptomatic case is of very low effectiveness, as the antigen profiles of the two individual infections are very likely to be too divergent. Further, if the viruses are significantly different and live virus is transferred, it risks presenting the recipient with two infections to deal with instead of just one.
-- 8th of 7, Apr 23 2020


Regarding changes in the virus:

//// Most mutations are deleterious, so mutants are less fit, and tend not to propagate well. This is not uniform over the genome.////
//This is entirely correct, and would be relevant if it were not for (a) the astronomic numbers of viruses released by an infected host, and (b) the very high transmissibility.
The genome is very small, and RNA based. Unlike DNA, there are no checks and balances in replication so the probability of a single bit error is high. 99.99% of those errors cripple the virus produced, and it goes no further. But since the numbers produced are so staggering, a few variations survive and thrive.</br> // A vaccine therefore ideally targets a part of the disease organism which can't change very much because of biological constraints. //
... except as [Sgt_T] points out, this constant shape-shifting is the exact mechanism the virus relies on. "Make many, many, poor quality copies, quite a few slightly different. Some will survive". And so they do.//

There are many different viruses, and they have very significantly different modes of attack. I suggest that it's a mistake to assume that one model will fit them all.
HIV functions in line with your description - but the process happens over years, within a single host.

Realistically, if covid19 was evolving at the rate you suggest, the disease wouldn't ever be cleared, and people would remain ill.
However, as I understand it (and remember, this isn't my specialism), the 'common cold' viruses (a group of which are somewhat similar to covid19) tend to cause a relatively weak (and shortlived) immune response. Nevertheless, the response builds up sufficiently to clear the disease. There are suggestions on the news that death from covid19 are caused by an overactive immune response -which is worrying but I don't think I can make any reliable comment on.

I think the critical factor you're missing in your model is that while some parts of the virus are relatively free to change, there are some parts which are much more constrained. (It's the external subset of these which are vaccine targets.)
You cite a value of 99.99% of errors 'crippling' the viron. I.e. that it cannot perform some essential function at all well.
I'm not sure where you obtained this value from, but (b) it's not a constant, and (a) it seems very high. In most biological systems, mutations are predominantly point-mutations- that is, they change one base. Within protein-coding regions, something under a third of single base changes don't affect the protein sequence, and are at worst somewhat deleterious. Obviously these don't affect any antigens, but they can open avenues for further mutations. A small fraction of point-mutations introduce a stop, which is likely fatal for the virus. But the remaining mutations change the protein sequence at exactly one position.
Anyway. For most proteins, most of the sequence just isn't that essential for function; at some positions almost any amino-acid will do, at some it needs to be an aa from a set with similar properties, while at others it could be one of only two or three. But there are some regions which are very important for function - active sites and binding clefts etc. Any changes there are typically really bad for protein function. These regions won't change very much, and if they're surface-exposed make good targets.
(If you do a protein alignment with a set of similar proteins from different species (or viral strains), some regions are conserved. That's what we're talking about.)
Can they mutate those sites to avoid detection /anyway/? Well, sure, but probably not all that much, and perhaps even then the virus will be a bit crap and easier to deal with anyway.
The other thing which can happen is a more significant acquisition of genetic information, perhaps from another virus. Obviously this could let the virus dodge some issue, but fortunately that's also a much, much rarer event.

Anyway. Rapid sequencing of viral genomes is a thing now. A quick bit of googling turned up a page which says:
"Based on current data, it seems as though SARS-CoV-2 mutates much more slowly than the seasonal flu. Specifically, SARS-CoV-2 seems to have a mutation rate of less than 25 mutations per year, whereas the seasonal flu has a mutation rate of almost 50 mutations per year."

Take that with a pinch of salt, but I'm not seeing anything which suggests a vaccine vs covid19 would be less feasible than one against flu.

I intend to return and reply to your other points after sleeping - it's late.
-- Loris, Apr 24 2020


We need something that can evolve faster and point to the most likely vectors, and then build our own dead virus set based on these. Complex predictive, (maybe neuro net?) or optimization software to the rescue, trained in all things basic-RNA viral. Yeah I know, that's probably too much math for even Watson to try and tackle, and builing it sounds like some seriously tricky work with CRISPR. But somewhere that way maybe the answer lies?
-- RayfordSteele, Apr 24 2020


That's an approach that's being tried; but see below. The behaviour is (mathematically) chaotic, so prediction is of necessity statistical, and preemption requires knowing exactly which complex proteins, out of near-infinite possibilities, to synthesze in bulk. Get it slightly wrong and you've wasted your effort, and worse, time, because time is the crucial resource here- the target is constantly moving, in unpredictable (chaotic) ways, because it's a biological system and not noticeably deterministic.

// and people would remain ill. //

They don't remain ill- the body reacts much faster than the virus changes, because it has a starting point - but they can be re-infected, because after 3 or 4 weeks it appears as a different bug and the immune system has to pretty much start over. It's not quite like malaria but not far off, in terms of symptoms, not mechanism.

// death from covid19 are caused by an overactive immune response -which is worrying but I don't think I can make any reliable comment on. //

Yes, that certainly happens. It's the primary mechanism for serious/lethal illness in the "young, fit" group. Their problem is caused by SIRS. A cytokine storm, as extensively mentioned elsewhere, is a SIRS phenomenon. It was raised as an issue on the HB on 23/03 in another thread. See also H1N1/Spanish flu.

// factor you're missing in your model //

Not ours - belongs to a Big Pharma corp.

// while some parts of the virus are relatively free to change, there are some parts which are much more constrained. (It's the external subset of these which are vaccine targets.) You cite a value of 99.99% of errors 'crippling' the viron. I.e. that it cannot perform some essential function at all well. //

The genome is very small. Each bit has to do a lot of work, and that perfectly ; unlike eukaryotes using DNA, there is no redundancy or error checking, no backup. An error can kill it dead, or stop it infecting, or allow it to still infect, but prevent it reproducing properly. Anywhere down the chain, it can fail.

A V8 will go on running with one failed spark plug. A single-pot motorbike just stops. Covid is a moped... but cheap and very plentiful.

// I'm not sure where you obtained this value from, but (b) it's not a constant, and (a) it seems very high //

No, it's not a constant, and it's cited as an indication, not a determined value, to emphasise that the vast majority of transcription errors are lethally damaging to such a small, fragile entity. See the analogy above re V8s.

// SARS-CoV-2 seems to have a mutation rate of less than 25 mutations per year, //

Yes, that's more or less what they've determned for a "full" mutation - a haf-life of 18 days, less than 7% match after 33 days. 365/18 is about 20 for 50% variance, 365/33 is about ten... so it's less than your data (which we don't dispute; for perfectly valid reasons, different researchers will get different results - this is biology, after all) thought their point is that after 18 days, on average the virus has changed enough in tems of expressed antigens for (a) tests to give a dangerously high proportion of false negatives, and (b) any vaccine to have such a relatively poor match as to lose effectiveness.

The further fly in the ointment is that the variance is a normal distribution. The mean divergence is 50% after 18 days BUT wthin that, some versions change hardly at all- the mutation doesn't affect the expressed antigens- and some vary wildly, and those are the real problem ones because they're just as virulent but look completely different.

But for a lot of not-well-understood reasons, Covid behaves very differently from influenza. One day, the best-guess pre-emptive tactics used to make flu vaccine may be adapted, but not yet, and not soon, no matter how hard the research is pushed by panicking politicos. It's going to take years, not weeks. By the time these mechanisms are understood (they're not insurmountable, it just takes time) the crisis will be over, and the virus will have killed everyone it's going to kill and gone on its way rejoicing, like all its brethren and sisteren since "life" evolved. There's still debate as to whether viruses are "alive" or just a blob of clever chemicals, but that's one for the philosophers.

<later>

// SARS-CoV-2 seems to have a mutation rate of less than 25 mutations per year, //

It may be that different researchers are interpreting "mutation" differently, so great caution is needed in comparing results.

One view is that mutation is "a change in the genome sufficient to cause a change in the antigen profile such that the virus is no longer recognized by a host with prior exposure"

Another view is that "it is a change in the genome such that the external appearance is unrecognizable but the virus can be unambiguously recognized by RNA profiling".

Both are valid uses of the term "mutation" - which after all is just a posh word for change - but have very different implications.

That could easily account for the perceived difference in "mutation rate".

Also, these changes are on a continuum. It's not like a nuclear decay sequence, where atoms undergo a very specific change and emit very specific particles (or photons). It's not like on the morning of Day 19 all the viruses are suddenly Variant 564 instead of 563. It's slow drift. Every day, new variations arise, some nearly identical, others very different. After your three weeks or so, on average there will be 50% divergence, but some specimens did all that on day one, and some won't change for months, if ever.

<later still>

After re-reading, the confusion over bacteriophages is explained. They have a highly efficient binding mechanism which allows them to attach to and attack bacteria despite quite large variations in the antigens those bacteria express. In essence, they have an adaptive, general-purpose clamp, which is pretty impressive for such a simple structure. Such an adaptive mechanism would be hugely useful as a therapeutic tool if it could be understood, but as yet it's not, at least not fully, and the functions can't be properly replicated by the big-molecule researchers.
-- 8th of 7, Apr 24 2020


Biology is still subject to physics. There are ways in which it attaches and ways in which it doesn’t, and that can be used to inform the analysis engine.
-- RayfordSteele, Apr 24 2020


//// I'm not sure where you obtained this value from, but (b) it's not a constant, and (a) it seems very high////

//No, it's not a constant, and it's cited as an indication, not a determined value, to emphasise that the vast majority of transcription errors are lethally damaging to such a small, fragile entity.//

Randomly choosing an unreasonable value is probably a mistake, even for rhetorical reasons.
A mutation can be lethal, or a range of deleterious, or about neutral, or a range of beneficial. Claiming that only one in ten-thousand mutations is not very deleterious or lethal is just wrong, because a significant fraction of most proteins could be changed (individually) without such a negative effect.

//// SARS-CoV-2 seems to have a mutation rate of less than 25 mutations per year, ////

//Yes, that's more or less what they've determned for a "full" mutation - a haf-life of 18 days, less than 7% match after 33 days. 365/18 is about 20 for 50% variance, 365/33 is about ten... so it's less than your data (which we don't dispute; for perfectly valid reasons, different researchers will get different results - this is biology, after all) thought their point is that after 18 days, on average the virus has changed enough in tems of expressed antigens for (a) tests to give a dangerously high proportion of false negatives, and (b) any vaccine to have such a relatively poor match as to lose effectiveness.//

As I think you realised later, that's a value over the entire viral genome. Obviously, for the purposes of a vaccine, to a first approximation only changes in the targeted antigen-encoding sequence(s) are significant, and that is a tiny fraction.
(I trimmed my comment for length - believe it or not- and a later paragraph on the web-page I quoted didn't make the cut. It stated that the covid19 genome is about twice as large as that of flu, so really the mutation rate is about a quarter of that. Perhaps leaving that in would have helped make that clearer.)
In any case, a sanity check which might have helped would be to consider that if each position of a genome were 'successfully' changing on average 25 times a year in a growing population, after a couple of months you would have every possible viable sequence present, and talking about covid19 would no longer confer useful information.

//It may be that different researchers are interpreting "mutation" differently, so great caution is needed in comparing results.//

It's not really that, it's that "a mutation" always has a context. If you're talking about a mutation in a genome, it can be anywhere. If you're talking about the mutation rate of a short sequence such as that encoding a surface-expressed loop, you're talking about a change in that specific region. Admittedly, this is often unclear in reports in the popular press.

//After re-reading, the confusion over bacteriophages is explained. They have a highly efficient binding mechanism which allows them to attach to and attack bacteria despite quite large variations in the antigens those bacteria express. In essence, they have an adaptive, general-purpose clamp, which is pretty impressive for such a simple structure. Such an adaptive mechanism would be hugely useful as a therapeutic tool if it could be understood, but as yet it's not, at least not fully, and the functions can't be properly replicated by the big- molecule researchers.//

There are many varied types of bacteriophage, with different approaches at each stage of their various life- cycles. (Viruses are arguably much more diverse than all other living organisms, since they use a wider variety of molecules for genetic information storage.) Some are more ecumenical than others in what they can infect. One important part of any virus life-cycle is how it gets into the cell - there are many strategies.

Also, you said "antigen" when you meant "receptor", or "binding site", or something.

(reply the other subtopics to follow..)
-- Loris, Apr 24 2020


Regarding issues with interventions in pregnancy:

The original comment was:
//A problem arises when extraordinary efforts "save" human infants that are not "100%". They grow to adults and reproduce, diluting the gene pool. In the long term, this weakens the species as a whole.//

//Large interventions, like in-utero surgery, can save a fetus that would otherwise have died, in anything other than a highly technologically advanced and wealthy society that can afford to spend the resource to do that. The question of whether it should be done remains unanswered.//

//One scenario is that as these interventions become possible, they also become more necessary to support "normal" reproduction. Fertility drops, more children need "help" to survive.//
...
//It is likely that, through the Law of Unintended Consequences and a perfectly understandable desire to "do good" for offspring, your species may slowly paint itself into a corner, genetically.//

To be honest, I'm not clear on what condition you're thinking of, because these between them cover a lot of ground. Your later comments reveal you're not talking about rare or sporadic genetic diseases (and which I think I dealt with).

So... humans are pretty crap at unassisted reproduction. As you observed, mortality rates in childbirth (both mother and child) have dropped massively as medical assistance has become available. And yet, there was still more than enough slack for the human population to expand massively before that; our ancestors were basically resource-limited. And at this point, it makes little sense to worry about fertility - there are more people than there have ever been. Arguably, too many.
But why is pregnancy and childbirth so hard? Well, as I understand it, the main issues are the bipedal stance and the brain size. Well, to be honest on reflection I'm entirely for interventions which increase the successful birth of more intelligent, large-brained individuals. In my view humans are onto a good thing with this thinking lark, and we should play to our strengths.

And natural selection can fuck right off with its teeth and claws; it's slow, it's imprecise and it's inefficient. At the point where we become able to safely improve our genomes to deal with all the shit, I'll be in favour.
-- Loris, Apr 25 2020


// after a couple of months you would have every possible viable sequence present, and talking about covid19 would no longer confer useful information. //

It doesn't now; the strains being examined in European labs are sufficiently different to the orginal strain categorized by the WHO as to be about ready for a new label all of their own. The Chinese-developed probes for the original strain don't work reliably on the new version, and it's nothing to do with sloppy procedures - it's a different beast.

// this is often unclear in reports in the popular press. //

It's pretty unclear in documents written by professional scientists, although they're not writing for a general audience. These aren't formal papers, more work-in-progress discussions.

And three new and very long documents have just arrived, one of which is turgid to the point of near-incomprehensibility. Asking for more information may prove in retrospect to have been unwise. We will attempt to distill a meaningful summary. The previous cut-paste-trim technique has, as you point out, created confusion as to what was meant by the original authors, who are consistent within their own work (i.e. "antigen" / "receptor"/ "binding site", etc. and "locus" seems to be used indiscriminately for multiple situations ) but not in the document over all, which is more of a concatenation.

<later>

O-Kay... should certainly have clarified previously that there is a significant difference between the overall mutation rate of the genome (as studied by one team), the mutation rate for loci critically involved in antigen expression (another guy) and the mutation rate of the antigens themselves (third group) all of which are inextricably interrelated, but different, and involve a mixture of hard sums and what looks very like finger-in- the-ar inspired guesswork.

// As I think you realised later, that's a value over the entire viral genome. Obviously, for the purposes of a vaccine, to a first approximation only changes in the targeted antigen-encoding sequence(s) are significant, and that is a tiny fraction. //

Yes, agreed. There are several "views" being presented on different portions of the problem within the same documnent, and they are hard to disentangle; some are contradictory as this is work in progress.

It may be some time before it is possible to present more information in a coherent form. The sheer volume of material is daunting.
-- 8th of 7, Apr 25 2020


//it's slow, it's imprecise and it's inefficient.//

That's the beauty, it catches the imperceptible subtleties that logic and design leap frog over to their engineered result. The only cost is wanting your children's children to have it rather than oneself.
-- wjt, Apr 25 2020


Ahh, altruism ...
-- 8th of 7, Apr 25 2020


//It doesn't now; the strains being examined in European labs are sufficiently different to the orginal strain categorized by the WHO as to be about ready for a new label all of their own. The Chinese-developed probes for the original strain don't work reliably on the new version, and it's nothing to do with sloppy procedures - it's a different beast.//

It does convey information, even if not as much as you're asking for.

Failing to get a hit on e.g. a PCR merely means that (at least) a single base has changed in the primers (if not that the specific reaction is a bit finicky). Slightly surprising, but not /necessarily/ a big deal. From my experience I would speculate that the issue is that the goal of the covid19-specific test was to actually be specific - so the primers chosen couldn't be in well-conserved areas (because then you'd give a false positives on 'common cold') ... and at the start of the outbreak there was very little information available to help guide the selection of good ranges.

But my point was, after two months of propagation, covid19 is still recognisably covid19, it's not diverged into every possible virus.

[on evolution] ///it's slow, it's imprecise and it's inefficient.////

//That's the beauty, it catches the imperceptible subtleties that logic and design leap frog over to their engineered result.//

No, natural selection really is sub-optimal as an algorithm. For example:

It doesn't (at least in higher eukaryotes) transfer successes. An innovation made in one species is unlikely to make its way to even moderately related species.

It loses stuff which isn't useful now but might be in future. For example, humans (and other anthropoid primates) need vitamin C in their diet because of a mutation in the synthetic pathway.

It can't fix legacy issues, even when they become significant. Because it effectively achieves improvement by trialling random bodges, there are some obvious sub-optimal designs which can't be fixed.

It's honestly sometimes surprising that it works as well as it does.

I hope that in the future, when we have sufficiently good genome-editing tools, we'll be able to make some pretty significant improvements over time. The above example, setting up a functional vitamin C synthesis pathway might be a good proof of concept. We'd make that change (and just that) in a few test cases, and see how that went. If those people have a problem... well, it's probably not too bad, and at least we could revert their children. If it's a success, we can roll it out to increasing swathes of the population.
And we could do hundreds of that sort of trial in parallel. Lots of the first batch would be to fix known genetic issues, like the trinucleotide repeats which undergo expansion to cause disease (like Huntington's, and fragile X syndrome). But after that, well, there's scope for plenty of increasingly ambitious developments as technology improves.
-- Loris, Apr 25 2020


// But my point was, after two months of propagation, covid19 is still recognisably covid19, it's not diverged into every possible virus. //

Two months in Europe/USA. It's been around in Asia for a lot longer.

We agree that // the goal of the covid19-specific test was to actually be specific - so the primers chosen couldn't be in well-conserved areas (because then you'd give a false positives on 'common cold') //

But a quick-and-easy blood test, not involving PCR, is more dependant on antigen structure, and those change a lot.

Trying to tease the simpler meaning the documentation, it suggests that all the loci that code for the surface antigens are clustered unusually close together. Thus any replication error in that region is likely to give rise to a variant without lethally affecting any other part of the genome.

We'll get back to you on that.
-- 8th of 7, Apr 25 2020



random, halfbakery