Natural Selection

Contents

Evolution by natural selection

Evolution is a change in gene frequencies in a gene pool.

Evolution occurs by natural selection, and by a number of other processes. We can divide evolution (rather arbitrarily) into microevolution (evolution within a lineage, not leading to speciation), which is largely explained by natural selection, and macroevolution (speciation events), which is also described by natural selection, but the processes of genetic drift and isolating mechanisms are very important to the outcome.

Speciation

Speciation requires some sort of isolating mechanism to allow two parts of a gene pool to diverge genetically. Sympatric speciation occurs when this isolating mechanism does not involve physical isolation. The incompatibilities of insect penes, reproductive isolation and sexual selection may play a role in sympatric speciation.

Allopatric speciation occurs when a real, physical barrier divides an existing species into groups that cannot freely communicate genetically (e.g. island chains, mountains, sea, river, rift valley). If the outlying population (that will eventually diverge into a new species) is formed from a small or non-random selection from the parent population, the founder effect may occur: this is a special case of genetic drift, the observation that small samples from a large population will not be completely representative of the parent species' genepool. Hence, small founding populations that have just made it over a mountain will probably be slightly different from the parent population. Likewise, low competition on the other side of the mountain, inbreeding due to the small population size, and release from stasis (?) may contribute to reproductively isolating changes.

Parapatric speciation involves fitness-related isolation based on a single gene locus. Heavy metal resistant plants may only differ at a single locus from susceptible plants (they may possess an advantageous malate pump, etc.), but homozygous suscepts without pumps and homozygous resistants with pumps do better in clean and polluted environments respectively, than the heterozygote does in either. This causes effective reproductive isolation.

Punctuated equilibrium is the idea that large populations are buffered against change (in stasis) for most of their history, and that change only occurs at speciation events, where both drift and selection contribute to the new species.

Doomed rivals to natural selection

Saltationism is the false idea that macromutations into unexplored areas of design space are responsible for adaptive evolution and speciation. Macromutations do occur, but are almost invariably harmful. The smaller a mutation is, the less likely it is to overshoot the optimum in one direction or another (like microscope focussing: if your microscope's nearly in focus, you don't wrench the knob to get it into better focus). However, macromutations of the 'Stretch DC-8' type may be adaptive and certainly must occur (no increase in complexity, just more of the same: e.g. snakes' vertebrae), but those of the Boeing 747 type (hurricane in a scrap yard producing a 747) almost certainly never are.

Mutationism is the false idea that mutations are non-random w.r.t. adaptation, i.e. mutations usually come up with improvements. This is demonstrably false. Mutations are non-random w.r.t. tissue type affected (skin vs. brain), type of mutagen (UV vs. chemical mutagens), gene locus, etc., but they are random w.r.t. adaptation.

Lamarckism is the (mostly) false idea that characters acquired during life are passed on to offspring. Particularly the idea that the striving of an organism will cause an organ to increase in size in the next generation. There are several problems with this:

Natural selection mechanism

Natural selection requires three things:

Natural selection feeds on variation, using it up. Several things keep variation in gene pools high (polymorphism):

The problem of blending inheritance: if inheritance is 'blending' then variation runs out as everyone becomes 'grey' at all loci. Mendelism (and the neo-Darwinian synthesis of Fisher et alia) solves this problem: genes are inherited discretely without 'blending'. The appearance of blending in many characters is due to a large number of genes, all with small effects on a single character. For an individual locus, one can only have black or white alleles, not any shade of grey.

Copying information of some sort or another is the basis of all life. DNA is a molecule that makes copies of itself: it is a replicator. However, replication is never perfect, and a replicator will occasionally produce copies that are not quite identical to itself, and differ in their power to get themselves replicated. Now, if you have a population of replicators, say DNA with different sequences, and you confine them so that they compete for resources, like the building blocks from which they are made, something interesting starts to happen. Some of the replicators will be better at replicating than others, and the numbers of each type of replicator in the population will change. That is, the population will evolve. Dawkins spells out the three fundamental things it takes to be a good replicator:

  1. Longevity. If you have a population of replicators, and some last longer than others before they lose the ability to replicate, then the ones that last longest will come to dominate the population. This is simply because if a replicator doesn't last a long time, then when we come to look at the population after a long time, it will not be there. DNA instructions are extremely long-lived: one group of proteins it makes are the histones, found in all eukaryotic organisms, which started to evolve from bacterium-like ancestors more than 2 billion years ago.
  2. Fidelity. If a replicator makes copies of itself, but very few of them are perfect, when we come to look at the population after a long time, the original replicator will have swamped itself with imperfect copies of itself. These imperfect copies are not the same as the original replicator, and as far as the census of the population goes, they do not count. So high fidelity replicators will come to dominate the population, because only they make copies good enough to be recognisable after a long time. The DNA-coded histone proteins from all eukaryotes are nearly identical.
  3. Fecundity. A replicator that makes many copies of itself will come to dominate the population (all things being equal) for the very simple reason that if a replicator makes more copies of itself than the competitors do, then there will be more of them after a given time. The histone genes are present in all eukaryotes, of which there are upwards of ten million species. As a conservative approximation, if each eukaryotic species were ten-celled and had a population of only one hundred thousand organisms, that means there are at least 10 000 000 000 000 copies of the gene in the world. Which is quite a lot.

These three statistical properties follow logically from the definition of a replicator. Evolution, which we define as a change in the relative proportions of a replicator in a replicator pool, needs only three ingredients: variation (e.g. as supplied by mutation), replication (i.e. heredity) and competition (i.e. some resource needed for replication is in limited supply).

Genes made from DNA, which are the replicators we're most familiar with, are long-lived, high-fidelity (digital, in fact) replicators capable of making enormous numbers of near-perfect copies of themselves. The ways in which they make more copies of themselves compared to their competitors can be quite bizarre and indirect, and involve such things as proteins, cells, whales and beehives, but fundamentally, replicators come to dominate their population if they fulfil these three criteria.

No matter how large the diversion (such as building an elephant) or how convoluted the route (DNA → RNA → enzyme → pheromone → action potential → behavioural change in another organism → mating → more offspring), these three things are what DNA is optimised for. A consequence of these things is that replicators do best if they benefit their own longevity/fecundity/fidelity, not that of others (unless in so doing, it benefits indirectly). Hence all genes are 'selfish', and 'altruistic' behaviours are seen to be selfishness in disguise: they are either reciprocal altruism (expecting a favour in return ) or kin altruism (benefiting copies of itself in close relatives).

The unit of selection is therefore often best considered to be the replicator (gene), as species and groups are always prone to exploitation by selfish individuals. Even individuals, which are merely temporary conglomerations of genes, are prone to exploitation by their genes, as evidenced by cancer, sexual selection, transposons, 'junk' DNA, and competition between male and female specific genes: a gene on the X chromosome for female fertility can cause some male infertility, and still be advantageous when averaged over the bodies in which it will find itself, which are twice as likely to be female.

…the antelope, knowing she is old and infirm, wills herself forward towards the advancing lion. Stumbling, and with tears in her eyes, she makes the ultimate sacrifice, throwing herself into the jaws of the cheetah, begging "eat me, eat me for the good of the herd!"…

You don't often see humans performing this sort of altruistic self-sacrifice. The chances of seeing an animal doing it are next to nil. As a rule, no sane animal (pets with bred-in kamikaze genes need not apply) will do something for the good of the herd, or the good of the species, or the good of the group. Happily, nature documentaries are catching up with this nonsense now, and have begun to expunge it from their scripts, so perhaps we'll see an end to this bad biology in the near future.

Competition and symbiosis

Competition may be:

Both types may lead to adaptation, and either may lead to evolutionary arms races, whereby two lineages will attempt to out-do each other: cheetahs run fast to catch prey, prey therefore have to run faster to escape, hence cheetahs have to run even faster, etc., etc. There may be some asymmetry in such races, and the life/dinner principle may apply: a cheetah is only running for her dinner, but a gazelle is running for his life.

Competition occurs because the Sun emits a constant flux of energy onto the earth, therefore primary production can only occur maximally at a rate approximating this flux. Organisms can reproduce exponentially, and consequently, the potential numbers of organisms will always outweigh the potential energy available for making such organisms, and competition for resources will ensue.

There are other relationships two organisms living in proximity to one another may have besides competition. These are generically called symbioses, and depending on the relative benefits to the two partner organisms (A and B), we can classify them as:

However it should be recognised that there is a continuum of variation in these categories. Peafowl exploit one another to rear offspring - the male has a silly tail and the female has to build the nest - but both have a vested interest in offspring. Such interactions cannot be easily pigeonholed. Remember, there are only replicators and vehicles! Biology is trying to bring order to the essentially unorderable ☺

Symbioses may occur at may levels, for example, between:

Symbioses also exhibit different degrees of physical proximity, here illustrated by parasitism:

There is little difference between predation and parasitism, besides (usually) the size difference between the 'partners'. It is really the mode of death that distinguishes them: death by ingestion is predation, death by eating from inside is parasitoidism (as in ichneumon wasps and caterpillars), and death by wasting is parasitism. Parasites need not actually kill their hosts, there is again a matter of a continuum. Some may be almost mutualistic, which is encouraged if the host and parasite share the same route to new generation: parasites that infect the germ line of a host (such as mitochondria and selfish DNA) have largely the same aim (survival-of-the-host) in 'mind' as the host itself. Some parasites cause slow death, generally if the parasite has some interest in host survival, e.g. if the parasite has a long life cycle and the host is rather rare. A fast death can be got away with if the parasite has no interest in host survival because there are many other potential hosts available and/or the parasite has a short generation time. This leads to the distinction between endemic and epidemic diseases: endemic diseases are those to which the host has some resistance and/or the parasite has found the optimal strategy to be the slow-burn approach. Epidemics are characterised by a large pool of susceptible hosts: the parasite can then be as virulent as it likes. Interestingly, many epidemic human diseases seem to have evolved from animal diseases (zoonoses): examples include HIV (SIV), measles (distemper, rinderpest), and flu (avian flu).

Here's a quick list of some interesting parasites for your delectation:

Game theory

The application of natural selection principles to the behaviours of unrelated organisms leads to some highly interesting results. Although altruism is never disinterested, nature may not be as red in tooth and claw as sometimes assumed, and organisms that have no genes in common (i.e. in which nepotism, or kin selection, does not operate) can still cooperate to mutual advantage. The canonical example from game theory useful to the discussion of natural selection is the prisoners' dilemma.

This is based on the idea that two prisoners are about to be convicted of a small crime. However, if one of the prisoners can be persuaded to turn Queen's Evidence on the other, he will be let off, and the other prisoner will be convicted on a worse charge. However, if both try to implicate the other, then both will be convicted of the worse crime, with a little leniency for confessing. Shopping your friend is called 'defecting' and keeping quiet is called 'cooperating'. My scores for this game are:

If you cooperate, then I should defect (+5 is better than +3). If you defect, then so should I (+1 is better than 0). Logically, then, we both conclude that whatever the other person's move, we should defect, and we both alway get punished. Doh! If only we could agree to keep quiet (and thereby always get +3, rather than +1)! Given that most social situations have an element of prisoners' dilemma about them, how on earth does cooperation ever happen, if it always pays (individually) to defect?

This is the realm of game theory, one of whose aims is to find the 'Nash equilibrium': where each player's response is an optimal response to the other player's strategies. In biology, the Nash equilibrium is called an 'Evolutionarily stable strategy' or ESS.

The iterated prisoners' dilemma: the solution to the problem appears to be iteration, that is, playing the game more than once. This provides the opportunity to build up trust between players. Players that cooperate with one another will both consistently get the reward, whereas a player that defects may be punished quickly and repeatedly by their partner. If the 'shadow of the future' is long (i.e. neither player knows how long the game will be running for), then strategies that allow both players to gain +3 may be found to be stable. Some of the more famous strategies include:

Eusociality

Eusociality is a special sort of kin-altruism characterised by a species having reproductive and sterile castes, the sterile castes acting as workers and facilitating the reproduction of the reproductive caste. It is found in bees, wasps and ants (Hymenoptera), termites (Isoptera) andtwo species of mole rat. The problem in explaining eusociality is explaining why workers that lay their own eggs do not do as well as those that help the queen raise hers.

The Hymenoptera are haplodiploid: males are haploid and females are diploid. The relatedness of two individuals is therefore:

The upshot is that sisters are more closely related to each other than mother and daughter, so on first glance, it seems that genes that make workers help make more workers via the queen should flourish. However, queen bees are polygamous, and most worker bees are half sisters (relatedness ¼), not full sisters, so this cannot be the answer. To look at it another way, a worker would prefer herself to the queen being the mother of the hive's new queens and drones. However, because most workers are half-sisters, she would still rather the queen made new queens and drones, rather than the other workers, because she is more closely related to the queen's offspring (particularly the drones), than she is to her half-sister's offspring. Hence, the queens' polygamy is an adaptation to manipulate workers into policing each other: any worker that does try to lay eggs will have her eggs eaten by 'jealous' half-sisters.

The Isoptera (termites) are not haplodiploid, but a similar manipulation goes on. Termite queens are long lived and outlive their kings. They then mate with one of their sons, who becomes the new king, and so on. This inbreeding leads to high relatedness amongst workers (who are of both sexes), and allows kin-selection to work in a similar manner to haplodiploids.

Naked mole rats are the only known non-insect eusocial species. Naked mole rats have soldier, worker, and reproductive castes, are of both sexes, and have long-lived, incestuous queens. They are basically termites with nipples, and the same arguments apply.

Origin of life

One of the fundamental issues we must decide before even embarking on such a discussion is to consider "what is life?". This is not as easy to define as we might like! Life has several properties, none of which are unique or defining, but which together contribute to our understanding of what a living thing is. Life is:

So when does the first evidence of improbable, information-containing, metabolic replication occur in the fossil record? The Earth is 4 500 million years old, as judged by several corroborating radionuclide studies of the oldest rocks on the planet show. Meteoric bombardment of the proto-Earth continued heavily until 4 000 MYA, probably precluding life during this period. The majority of the oldest rocks on Earth are 3 500 million years old, and the earliest microfossils are from 3 000+ MYA, hence we only have a window of about 500 million years from the end of the meteoric bombardment to the first signs of microbial life. This means we are either very lucky, or life is a nigh-on certainty!

No-one knows where life originated. Some believe it came from space in some way, however, this 'panspermia' just shifts the blame from Earth to elsewhere, and isn't particularly helpful (we will Occam it away for the purposes of this essay). Looking to possible terrestrial origins, we find a wealth of theories, including ponds (but these were probably rather ephemeral, especially during the tail-end of the meteoric bombardment), the sea (but this is probably too dilute), ice bound water in e.g. frozen ponds (but reactions are very slow here) and hydrothermal vents (but the temperature varies between 20 and 1200°C in the space of a few metres). None of these is knock-down convincing, but hydrothermal vents are the current favourite.

What did life use as its raw materials? Early theories of life's origins thought that the earth had a reducing atmosphere (i.e. lots of ammonia and methane) but this seems less likely now, as our understanding of early Earth chemistry has proceeded. An oxidising atmosphere (oxygen) came very much later as evidenced by absence of rust in earliest rocks (the cyanobacteria were responsible for this bit of environmental vandalism), and it's likely that the original atmosphere of Earth was very dull and fairly neutral (nitrogen, carbon dioxide). This is one of the reasons that hydrothermal vents have become popular: life involves highly reduced carbon compounds, and the only places reducing agents are in abiotic abundance are in places like vents where gases escape from the Earth's mantle. In addition to the contributions from the atmosphere (or lack thereof), meteoric waste (amino acids, water-ice, cyanide, polyaromatics) and hydrothermal-associated chemicals (hydrogen sulfide, carbon monoxide, cyanide, pyrite) may have contributed to the alleged 'soup' from which life evolved.

When discussing the first organisms, we should distinguish between the most recent common ancestor of life (which may, or may not, have been Archaea-like) and the first forms of life. These may not necessarily be the same thing: imagine all life but mammals were wiped out. We might start invoking something shrew-like as the first organism, but this is clearly ridiculous. Similarly, DNA/protein creatures may just be lucky survivors from a far more diverse group of proto-life. However, bearing this in mind, we note that the Archaea (those weird-arse bacteria that live in boiling sulfuric acids, etc.) are sulfur metabolisers and hyperthermophiles, supporting a hydrothermal origin, if indeed these features in Archaea are 'primitive', of which we have no guarantee.

The chicken and the egg

Life has two main biochemical features: metabolism (proteins, and hence complex biochemicals); and replication (RNA/DNA, and hence the capacity for evolutionary improvement). Unfortunately this drops us straight into a chicken/egg paradox: RNA/DNA requires complex biochemistry to produce the bits from which they are formed, hence proteins must have come first. But proteins cannot be improved by replicative natural selection, because they do not self-replicate, so RNA/DNA must have come first. Oh dear.

This has led some to consider whether there some 'scaffolding' over which an 'arch' of proteins and RNA were built, that no longer exists, leaving the arch free standing, and us none the wiser as to how the synergy of RNA and protein came to be. Some have suggested a precursor replicator that came before RNA, possibly clays, which appear to have some rudimentary ability to replicate their crystal form and manipulate their environment (i.e. they have both a rudimentary phonotype and genotype). Precursors to metabolism (i.e. complex chemistry) can possibly be achieved on the surfaces of catalytic clays.

Metabolism requires an energy source. Light (phototrophy) seems unlikely in this capacity, as white light is non-reducing, but UV, which is reducing, is usually too ionising, and neither are available in the favourite haunts of origin-of-life theories (vents, sea-bed, etc.). The concept of a primaeval soup (organotrophy), also seems unlikely, as this soup (the sea and whatever meteorites and reactions in the atmosphere had deposited into it) would be a very dilute and rather dull, biochemically. Chemical energy (lithotrophy) seems rather more likely, in hydrothermal vents at least, where rather exciting chemicals are found.

There are several groups of molecules whose prebiotic synthesis we should be interested in. Perhaps the simplest are lipids (fats): these are used today as energy stores, and for the compartmentalisation and concentration of chemicals within a cell. However, short branchy lipid chains are more probable in prebiotic chemistry (without complex enzymic catalysis) than long straight chains, which are much more useful for forming bilayers, etc. They are also energetically expensive and have solubility issues. The porphyrins (and relatives, such as vitamin B12) are another group of lipids, used in electron/oxygen/light interactions in cells in systems like cytochromes, haemoglobin, phaeophytin, chlorophylls and phytochrome, and are also presumably very ancient.

The most important group of chemicals for prebiotic synthesis to create are the nucleotides, the building blocks of DNA and RNA. These are used for data storage (but DNA is more stable than RNA), and also in catalysis: the ribosome, which directs protein synthesis in all cells is a gigantic catalytic RNA complex, or ribozyme. Tiny pieces of RNA and its relatives are found in many enzyme cofactors (like FAD, NAD, ATP, CoA and many vitamins), which may give hints as to the sorts of compounds that were originally involved in prebiotic metabolism: maybe the bits of RNA in today's proteins are the remnant of an earlier age when most of an enzyme was RNA, with bits of protein as cofactors. There are in fact many 'hidden' ribozymes in the cell: enzymes that actually do what they do using a piece of catalytic RNA. These include ribonuclease-A, snRNAs, tRNA, rRNA, self-splicing introns and telomerase: it's noteworthy that these are all involved in extremely basic functions of the cell, such as translation, transcription, replication and splicing. Unfortunately, deoxyribose and the pyrimidines of nucleotides are extremely difficult to form abiotically. Purines can be made by the electrolysis of cyanide, as might occur in the atmosphere, and ribose can be made by the formose reaction from formaldehyde (although in extremely low yield). Our understanding of the formation of nucleic acids is not much better than that for lipids! RNA seems a more likely precursor replicator than DNA, because it seems to combine the roles of replicator and catalyst in one molecule, however there are problems with this idea, not least of which is RNA's chemical instability. Some have proposed that peptide nucleic acid (with a backbone of peptide-linked N-(2-aminoethyl)-glycine rather than sugar-phosphate), which is far more robust, may be a useful starting point for the discussion of proto-RNA.

The majority of organic molecules are neither RNA, nor even protein: they're the 'dull' carbohydrates and organic acids of intermediary metabolism, energy storage, and polymers like starch (energy store) and cellulose (structural). These can be formed by formaldehyde condensation, but because of the number of chiral centres in carbohydrates, there are severe problems in producing any particular stereoisomeric form in any reasonable yield by abiotic processes.

The final important group to consider are the proteins, peptides and proteinoids. Amino acids can form from simple chemistry under 'reducing' conditions (such as in hydrothermal vents full of hydrogen sulfide). However, their condensation to form proteins is difficult in aqueous solution, because of the competing backwards reaction of hydrolysis. Proteins are extremely important to life, as they are needed for both catalytic and structural roles, and interact with other cofactors such as bits of RNA and metal ions. Proteinoids may be a precursor of today's protein: conventional proteins are formed purely by condensation to form one sort of peptide bond. Proteinoids on the other hand are condensed together in any way they chemically can.

Specific theories of the origin of life

These may be broadly divided into "metabolism first" (MF) and "replication first" (RF) theories.

Oparin's theory (MF). Lightening in a reducing atmosphere yields purine, alanine and glycine, and colloidal systems mimic 'protoplasm'. Metabolism was based on proteinoids, but how can these evolve without a replicator? Oparin was trying to do the best he could at the time, but without explaining RNA, the theory is too simplistic.

Haldane's contribution (MF). Haldane recognised the importance of ozone and UV interactions (good and bad) with prebiotic systems.

Dyson's stochastic evolution (MF). Compartments of autocatalytic proteinoids contain information. Evolution can occur with such systems as long as there is some heredity: e.g. if offspring compartments are similar in biochemistry to their parents. If compartments take up some precursors from the environment more than others, this will encourage certain proteinoid syntheses. This is an interesting take on trying to reconcile the metabolism first theory with a basic sort of replication and evolution.

Kaufmann's self-organising complexity (MF). In a big enough pool of proteinoids, there must be an autocatalytic cycle, i.e. a set of proteinoids that is able to form itself: A makes B, B makes C and C makes A. However, mutations break these cycles, and the mathematical models are somewhat dodgy.

Morowitz's lipid centric theory (MF). Morowitz considered micelles, vesicles and cells to be the most important 'inventions', preceeding both replication and true metabolism, which came later, via mechanisms like the influx of precursors via selectively permeable membranes, and electrically charged chromophores in membranes yielding biochemistry. However, micelles are more readily formed than vesicles, but cannot support very exciting chemistry within their lipid core, and the theory fails to address where the rest of biochemistry comes from.

Spielgelmann and Orgel's RNA studies (RF). Q-beta replicase (a nucleic acid polymerase), an RNA template and a supply of nucleotides can undergo selection for fast replication in a test tube, yielding a 'new species' of truncated RNA. Orgel found that small pieces of RNA and nucleotides will replicate even without a protein catalyst, but slowly, inaccurately and restrictively.

Eigen's RNA world (RF). Eigen further found that QB replicase will make RNA from nucleotides without a template at all! He thought that ribozymes (RNA catalysts that can also be replicated) formed a prebiotic 'RNA world' of RNA hypercycles. Much like proteinoid hypercyles, RNA hypercycles are of the form RNA-1 helps replicate RNA-2 via a proteinoid (P1) it codes for. RNA-2's proteinoid (P2) in turn helps replicate RNA-3, whose proteinoid, P3, helps replicate RNA-4, which closes the loop by helping replicate RNA-1 via P4.

A hypercycle, as described in the text.
A hypercycle. The RNAs are self-replicating RNA molecules, each also producing a proteinoid enzyme (P) that aids the replication of the next RNA in the cycle.

However, this still requires some primitive metabolism in the early stages to form the RNA: proteinoid may have acted as replicases before RNA replicase ribozymes. Autocatalytic hypercycles of replication could be compartmentalised by membranes to form 'quasi-species'. The biggest problem with the theory is that of information meltdown. RNA > 100 bases long will neither replicate accurately nor be formed spontaneously with any likelihood. RNA < 100 bases long will have no catalytic activity of note. RNA is also highly unstable, and if we rely on spontaneous generation of the first self-replicase, we require two RNA replicase molecules to arise spontaneously for it to work, which is astronomically unlikely!

Quasi-nucleates (RF). Why stick with RNA and proteins? We can also consider amino acid/RNA hybrids (like PNA), xanthine bases substituting 'normal' bases, hexose sugars or threose in nucleic acids, and DNA for our original replicator.

Cairns-Smith's clays (MF). Why even stick with RNA? Why not have clay as the original replicator and catalyst?

Wächtershäuser's pyrite metabolism (MF). Autotrophic thermophilic sulfur metabolism as the original metabolic system. This theory has much to recommend it, particularly after Woese's discovery of hydrothermal vent Archaea: vents often form pyrite. Wächtershäuser's metabolism is based around the reaction:

FeS + H2S → FeS2 (pyrite) + H2 (reducing power) + energy.

The positive charge on the surface of pyrite binds negative species (amino acids, nucleates), which exerts a selection pressure for certain sorts of metabolism and replicator. This surface chemistry solves the dilution and hydrolysis problems mentioned earlier. Hydrogen can be used to fix carbon dioxide to yield water and alkyl chains, which can form lipids, used (maybe with sulfide membranes) to compartmentalise off cells: an accumulation of pyrite causes 'hemicell' division, and eventually sloughing of cells. Like Dyson, we have a stochastic evolution (but maybe a Cairns-Smith's clay replicator is possible too). A Krebs-like cycle (reductive citrate cycle) can be made to run on pyrite in the presence of carbon oxides, yielding succinate. Thiols and thioesters also form, as do thiocarboxylates (q.v. activated acetyl group bonded to mercaptoethanol in coenzyme-A). However, amino acids don't seem to form, although thioesters can help cross link proteinoids.

In summary, we certainly don't know what the origin of life is with any certainty, but the above discussion gives you some idea of how far we have come, and how much closer we are to understanding where we ultimately came from.

Test yourself

  1. What is evolution? How does it differ from natural selection?
  2. What is the most important factor in promoting speciation?
  3. Why are mutationism and saltationism incapable of accounting for adaptations?
  4. What are the three requirements for evolution by natural selection? How are they met by living organisms?
  5. How might polymorphism in the banding patterns of snails be kept high by predation?
  6. What are the three requirements of a successful selfish replicator?
  7. How do mutualism, parasitism and predation differ?
  8. Why might germ-line parasites undergo selection pressure to become mutualistic?
  9. How can 'nice guys finish first' in an iterated prisoners' dilemma?
  10. Why is there no selection pressure on a worker ant to produce offspring of her own?
  11. Discuss some of the problems associated with investigating the origin of life.

Answers

Peer Review
This page has been peer reviewed by 3 people. Thanks to Selma Koghee-van Ommen and Lotta Sundström for their feedback.