Thread by @TRyanGregory

Evolution 101. 🧵

Mutations occur as chance errors in replication. They're just mistakes in copying. Most have no effect. Some are detrimental to the organism (or virus), a few may happen to be beneficial -- this depends on the environment.

Three main evolutionary processes determine what happens to genetic variation once it arises (and these are independent of the process that generates new variation, namely mutation): genetic drift, natural selection, and gene flow.

We call different versions of a gene "alleles" and we can talk about the proportion of those versions in a population as "allele frequencies". Genetic drift is a random change in allele frequencies that occurs by chance.

Genetic drift is basically sampling error, in which the genetic variation in a new generation does not accurately reflect what was present in the previous generation. The most common mechanisms of genetic drift are founder effects and population bottlenecks.

Founder effects occur is a random, non-representative subset of a population moves to a new location and founds a new population. The allele frequencies in that founder population won't be the same as in the source population, and there will typically be less variation overall.

Population bottlenecks are sudden, severe declines in population size in which survival happens at random. So, for example, a drought or storm causes a major die-off and the individuals who survive were just lucky (rather than having traits that helped them survive).

Gene flow is movement of genetic variation from one population to another. The overall effect is to introduce new variation into an existing population (if the source population has different alleles) and to make two populations that exchange alleles more similar to each other.

When we think about the early stages of new species evolving, we generally are considering ways that gene flow is being blocked. Lots of gene flow means two populations are less likely to diverge genetically.

The final evolutionary mechanism is natural selection. In this case, the reason some individuals survive and reproduce better than others is *non-random*. It doesn't occur by chance. It is specifically related to heritable traits that make survival and reproduction more likely.

This is where the concept of "fitness" is relevant. In evolutionary terms, fitness refers to the advantage in survival and/or reproduction due to heritable traits. Fitness depends on the environment. What is fit in one environment may be neutral or unfit in another environment.

There are different forms of natural selection, depending on what part of the distribution of traits is fit/unfit in the population: directional selection, diversifying (or disruptive) selection, and stabilizing selection.

Under directional selection, one extreme of the distribution is fit and the other is unfit. This drives the distribution of traits in a particular direction from one generation to the next.

For example, if the largest individuals leave more offspring on average than smaller individuals in each generation, then the average size will increase over time. Not because individuals start being born larger in response, but because more offspring are born of large parents.

In diversifying selection, it is the two extremes that are fit and the average traits that are unfit. This can cause a population to split into two. For example, if the smallest and largest individuals do well but the medium-sized are at a disadvantage.

Finally, in stabilizing selection the average value is fit and the extremes are both unfit. The result is that this prevents the distribution of traits from changing in the population because deviations from the current average are detrimental in that environment.

Here is an image from Wikipedia to show these three types of natural selection. (A) is the original distribution of traits, (B) is the new distribution. 1) Directional, 2) Stabilizing, 3) Diversifying selection. en.wikipedia.org/wiki/Natural_s…

One other form of selection is called balancing selection. Balancing selection maintains variation in the population whereas the others deplete variation (again, the source of new variation is random mutation -- selection does not cause new variation to arise).

Balancing selection can happen through heterozygote advantage or frequency-dependent selection, but that's more complicated than this basic overview.

So, to review. Mutation is the source of new variation. It happens as random errors in replication and is independent of what happens to variation once it exists. Genetic drift is random changes in allele frequencies due to sampling error.

Gene flow is movement of alleles among populations and makes populations more alike. Natural selection is non-random changes in allele frequencies resulting from differences in survival and reproduction because of heritable traits that are fit/unfit in a particular environment.

There are several forms of natural selection. Some cause change in the distribution of traits in a population (directional, diversifying), some keep a population from changing (stabilizing).

There are many factors that affect these evolutionary mechanisms. Mutation rates can be high if there is weak quality control and repair of errors, or if there is some environmental factor (mutagen) that messes up replication.

It also matters how much replication is happening. Every time a genome is replicated, there can be errors. Lots of replication means lots of opportunities for mistakes to occur. (In multicellular organisms, only mutations in the germline are relevant in evolution, of course).

As to what happens to alleles, this depends on the environment as well as population size. Genetic drift, which is sampling error, is stronger when samples (i.e., populations) are smaller. Natural selection, which is non-random, is stronger when populations are large.

Whether an allele is fit or unfit (will be subject to non-random natural selection) or neutral (will evolve by random genetic drift) depends on the environment.

Natural selection and genetic drift can happen at multiple levels. The main one is, of course, among organisms within populations, but these can also happen within organisms. Cancer is an example of cell-level selection that is usually suppressed in multicellular organisms.

When it comes to viruses, there are two levels as well: within hosts and among hosts. Because viruses mutate so quickly (by chance, because their repair mechanisms are weak), there can be new variation arising within a single host.

Some mutants will do better within the host -- that is, they will be better at invading host cells or will be replicated more quickly than other versions of the virus within a host. So, there is natural selection within the host.

Some mutants will do better at getting into new hosts. For example, maybe they form smaller aerosol particles and spread father when sneezed out. Or maybe they are in high concentration in the nose rather than deeper in the lungs, so they get shed more easily.

The mutant viruses that do best within a host are not necessarily the same ones that do better at infecting new hosts. In fact, a highly virulent version might be very effective at invading host cells but do so much damage that the host never spreads it to another host.

There can be a trade-off between virulence (replication within a host that causes damage to the host) and transmissibility (spread to new hosts).

Which versions of a virus evolve depends on the mutations that happen to occur by chance replication errors and the outcome of genetic drift and natural selection both within hosts and among hosts.

Whether viral evolution involves increased or decreased virulence and/or higher or lower transmissibility depends on many factors. Number of replication events happening. Rates of replication errors. Selective pressures within and among hosts. Viral and host population sizes.

Virulence and transmissibility are not the same thing, and there may be trade-offs between them, but it's also a concern that a virulent (damages or kills the host) virus can still be successful at the host population level if it is able to spread to many new hosts.

Viruses that are both highly virulent and transmissible will eventually run out of hosts to infect, but they can do great damage before that happens.

One of the many positive effects of reducing transmission (e.g., with vaccines, masks, etc.) is that this imposes a selection pressure for less virulence. If only versions of the virus that don't incapacitate or kill the host manage to reach new hosts, then those are fitter.

Reduced transmission also means fewer replication events happening and this means fewer new mutations.

A mild but highly transmissible version of a virus can spread quickly through a population and then fizzle out as hosts become immune, and many people seem to be assuming this will happen with Omicron, but that also means a lot of replication and new mutations.

The Omicron variant in particular has many, many mutations specifically in the spike protein, which is one reason it is so much more transmissible and escapes previous immunity. And this may now be the starting point for new variants.

It is possible that Omicron is milder (than Delta, at least) and that it will infect pretty much everyone and that this will be a step toward SARS-CoV-2 becoming endemic (like flu, requiring seasonal vaccinations).

But it is also possible that Omicron may undergo more chance mutations that make it more virulent as well as highly transmissible. Then it spreading rapidly will mean many hospitalizations and deaths before it runs out of hosts. We do know it is still evolving.

Viruses don't want anything. They just spread to new hosts or they don't, and replicate effectively in hosts or they don't. Mutation, genetic drift, gene flow, natural selection. There are many factors we can't control, but there are some that we can. We really ought to try.