Rhododendron walks into a bar...

By Anne Buchanan and Ken Weiss

Plants may be stuck in one place, but they aren't just sitting there twiddling their thumbs.  That they respond to environmental cues has long been known, of course, and the hormonal and molecular mechanisms for responding to light, temperature, moisture and so forth are well-established.  Some responses evoked by one part of a plant, such as attack by predator, can communicate to other parts or even to nearby plants.  But it has only fairly recently been suggested that plants are also able to recognize kin, and respond differentially, and in ways that enhance reproductive fitness, to the presence of plants to which they are closely related.

Among other reported indications of kin recognition, plant roots have been found to grow more in the presence of 'strangers' than 'kin'; kin recognition is said to be via root-derived cues, though that this actually happens is not without controversy, primarily because the molecular mechanism has not been identified (from 'Shedding light on kin recognition response in plants,' New Phytologist, Bias, 26 Nov 2014).  But now a new paper ('Photoreceptor kin-recognition among plants,' New Phytologist, Crepy and Casal, 29 Sept 2014) describes a possible mechanism for kin recognition among Arabidopsis thaliana, or mustard weed, the most frequently studied model plant.

Crepy and Casal did a series of experiments growing Arabidopsis plants in pots, to eliminate the possibility of confounding cross-talk between root systems, in variations on the theme of single genotype or mixed genotype rows, with plants either surrounded by kin or non-kin.  They also included mutant plants with known responses to different light waves, and plants of different ages, all exposed to differently filtered light sources.  They were interested in whether there were effects of proximity to kin, which they measured in terms of how close leaves were to neighboring leaves, or how much light fell on each leaf.  The idea was that kin don't compete over pools of light, but instead allow their relatives equal access.

Bias writes:

Crepy and Casal showed that plants recognized their kin neighbors by horizontally reorienting leaf growth compared with the interactions with the nonkin members. The authors also showed that the mechanism that led to reorientation of the leaf with kin members was regulated by phytochrome B and cryptochrome 1. The work by Crepy and Casal provides the first molecular evidence of the way in which plants respond to kinship.

They also showed that plants that interacted with kin produced more seeds than plants growing among strangers, "a clear indication of mutual benefit and cooperation."

From Bias, 2014;
'Aboveground and belowground interactions in plants experiencing kin and nonkin members.'

Have Crepy and Casal demonstrated beyond doubt that plants recognize kin?  Probably not; it has been controversial, and will surely remain so.  For one thing, there needs to be a convincing mechanism for recognizing what 'kin' means, and this is a serious issue both practically and theoretically.  In animals, with their various pheromones and highly variable immune/identity systems, the latter of which can be highly variable because mutations accumulate rapidly, various receptors and detection systems, also part of the system whose function is in mating or immune defenses, recognition of close molecular similarity in these aspects of the genome would be called 'kin'.   Plants also have high-variability immune-like systems for detecting invaders but whether they monitor this for self or self-like patterns is something we, at least, don't know.

It will take much beyond this rather limited study before any serious evolutionary geneticist will be completely convinced.  This is, in part, because authors must show beyond a reasonable doubt a molecular means specific enough to detect and evaluate the degree of kinship rather than just same-species or locally same-environment, both of which could affect many aspects of plant molecular biology.  Can a plant tell a clone from a cousin, say?  Finding such evidence has proven generally to be a very tall order, but of course that doesn't mean the Crepy and Casal finding is wrong, but it does need to be viewed with circumspection until details are known, because empirical findings like theirs can have multiple explanations.  The reason is not hard to see, and it seems there are some semantics involved, with the meaning of the term 'kin'.

The basic idea and rationale of kin selection
The main underlying idea that makes this of any interest has to do with altruism.  Helping any other organism may be at your own expense, and put you at a reproductive (and hence evolutionary) disadvantage if, say, it costs energy to help the recipient gain resources that lead to its greater reproduction when you could be putting that energy toward your own reproduction.  If that's the case, the genetic variants that lead you to do this won't proliferate as much as the recipient's. The explanation offered mathematically by William Hamilton over a half-century ago was that if your aid to a relative of a given degree -- and this is where 'kin' comes in -- must lead the recipient to reproduce more by a factor at least as great as your direct kinship relationship in order for the behavior to evolve.  In animal terms, you share half your genome with your sibling.  If you lose an offspring because you helped your sib, s/he must produce more than 2 additional offspring as a result of the help -- that is, in the next generation (on average) there will be at least as many copies of the altruism-inducing variant. If the recipient is of a more distant degree of relationship, the advantage must be much greater than just 2 for 1. 

Hamilton's rule was for decades a kind of cult religion among strong evolutionary determinists looking for precise natural selection everywhere.  To be fair, it also was a response to accounting for the evolution of what seemed like self-defeating behavior, to counter a heretical argument that invoked 'group selection', that organisms could evolve behavior that was self-limiting if it was good for the group.  This was heresy in the sense that it went against the rugged individualism of arch-Darwinism -- and mathematical analysis showed that was a more problematic phenomenon to account for.

However, careful quantitative ecological genetic studies have generally not supported the idea as of much practical applicability except in unusual circumstances.  On the other hand, in general if you help another member of your species, relative to other species, or if because you drop your seeds near where you live, your neighbors are  your relatives, then such behavior is easier to understand and doesn't require great precision -- for example, you don't have to have a mechanism for genotyping your neighbor, as you effectively do under Hamilton's rule.  If your neighbors are your relatives, helping is OK, and this can be so even if it's just within species if you are (or your ancestors at the time the helping mechanism evolved were) locally reproducing.

Likewise, if you compete for soil nutrients or sunshine with other plants where you live, it can in principle at least (this needs to be shown quantitatively) mean that you're better off by helping your species members rather than some other species.  Again, in the Hamilton's rule sense they share your genes far more than other species do.  These sorts of things can in principle account for the kin-recognition in plants that this paper refers to: it does not have to be testably close kin.


In this case, we think that the Crepy and Casal idea seems to have rather misleadingly used the term 'kin', not to refer to close family relationship of known degree but to what amounts to more distant group relationship.  Distant groups should not be referred to as 'kin' in this kind of situation because its connotation can be unclear.  After some variation has accumulated, it is clearly reasonable to ask whether similar molecular physiology might induce similar responses, or cross-reactivity, in those from the same group compared to those from a distant group.  No evolutionary kin-rule selection of the precise Hamiltonian kind need be involved.  In the wild, ancestrally, neighbors are kin.  In this case, if the story holds up, one will want to know how these particular genes effect such cross-reactivity.

Is it a perfect good?
But let's say that Crepy and Casal have demonstrated that plants help their kin.  And that kin recognition is an unalloyed good, with demonstrable fitness benefits.  As they point out,

Preferential helping of relatives has been observed for a wide range of taxa. For instance, in vertebrate (bird, mammal) species, helpers preferentially aid closer relatives during breeding (Griffin and West, 2003). In the social amoebae Dictyostelium discoideum, cells cooperate preferentially with relatives and aggregate to form multicellular fruiting bodies (Hirose et al., 2011). In humans, as the cost of helping increases, the share of help given to kin increases, whereas that given to nonkin decreases (Stewart-Williams, 2007).

The assumption is that the more closely related organisms are, the more likely they are to cooperate.  That is, increasing the reproductive fitness of one's kin is good for one's own fitness.  The reason that kin selection has been, and continues to be a hot topic in human evolution is both that it confirms hyper-darwinian fine-tuned selection, which many hold as fervently as a religion, and that it accounts for cooperation without being culturally wishy-washy (as they'd see it), and that there should be a mechanism to account for its evolution; note of course, that mechanism needs to be quite specific to define 'kin', which cannot just be assumed (as we outlined in the previous section).  But we, at least, have plausible molecular-genetic means of detecting close kin.

But people aren't plants, or dictyostelium.  We also have culture, and it can powerfully affect our behavior, so that what makes sense evolutionarily for other organisms doesn't always apply to us.   For one thing, culture allows us to assign relationships symbolically rather than just genetically.  We can imagine what 'success' may mean (e.g., getting to heaven), which goes beyond mere Darwinian proliferation.  We can form clans or other structures based on all sorts of criteria, not just genetic relationships. We don't always optimize our own fitness.  Humans are the only organisms that abort their own fetuses, that blow themselves up in support of an ideal, that have civil wars, killing people in fact most closely related to them. But humans also rescue strangers from drowning, and grow food to be consumed by people across the world.

We can devise all the equations we want about kin and fitness, and we can calculate the heritability of this trait and that, to show that behavior is genetically determined, but our thinking brains, and the power of culture trump those rules.  Members of Homo sapiens simply are not just bags of 'selfish' genes, and meaning isn't just reproduction.  We're not nearly as hard-wired for behavior as many other 'lesser' species of animals and plants are often (correctly or incorrectly) assumed to be.

On the mythology of natural selection. Part IX: What would it mean if selection really were deterministic?

Today's post will have to do with the question, "What is cause?". This will be a bit abstract, but I hope it will be at least somewhat understandable.

We have been writing about the subject of genetic determinism and its twin, deterministic natural selection.  Pure determinism means that when I know an organism's genotype I can perfectly predict its traits and that given different organisms' genotypes I can perfectly predict their relative fitness (reproductive success).  It's one thing to consider skylarks and oak trees in this regard, because nobody really cares; but now we know that humans are part of Nature rather than separate creations, we naturally want to apply the same principles to ourselves as we do to plants and other animals.  And though issues about determinism apply to all of Nature and its evolution, it's the way they hit home that makes it important for the human sciences, above all, to realize the implications of how we view ourselves in light of our ideas about evolution.

Some people object to suggestions that natural selection might not be genetically deterministic--that there might be other aspects to the organisms' traits, such as environment which in the case of humans includes their culture.  An unstated or perhaps even not even fully aware objection to any challenge to genetic determinism is that if you open the door, even a tiny crack, to the uncertainties of probabilism or environmental, or in the case of humans to cultural effects, you'll venture off into 'softer' non-scientific views of Nature.   That point of view suggests that we can peer into genomes to see individuals' or groups' inherent worth, as one can fancy that Nature in the form of selection sees and has seen it.

To use humans further to illustrate the point, the fact that we wear coats in winter and domesticate our food sources would seem to be manifestly obvious examples of niche construction.  To push this further, can any sane scientist argue that, say, Doric architectural columns were built because of specific genotypic differences between their architects and the architects who built Ionic columns?  If not, and if it's not cultural (environmental), what is the 'cause' of the style differences?  An honest assessment raises the question how far can assertions of genetic determinism go, and if the door is opened to cultural/environmental causation, where do you stop, if you can't read individuals' traits off their genomes?

Doric Order of the Parthenon (Wikimedia)

Ionic Order (Wikimedia)

But let's think a bit deeper than this superficial bit of sociocultural generalizing.

If evolutionary fitness were truly determined
In this series on natural selection we've often accepted, temporarily for the sake of argument, a proposition that genetic causation, and hence selection really are perfectly deterministic.  That means selection reads two organisms' genotypes, say, and determines everything about them, including who will win the evolutionary race, in the same law-like way that gravity pulls objects towards each other as a deterministic force.  Let's again take this as true for the purposes of argument, and see its consequences.

First, such determinism if true would mean that everything--everything--about every organism is perfectly predictable.  And if what happens is perfectly determined and predictable, then we can follow what's here today, using the force-like laws that make it predictable, back as well as forward in time, so that in fact everything--everything--is just following those laws and hence was already predictably inevitable from the instant of the Big Bang.  Of course, to actually predict everything you'd have to have all the information, perfectly understood, which in turn requires that you be above or out of the system whose properties you want to know: in a sense you need to know more than everything there is to know.  Arguing that position is logically no different from just saying "God willed it to be that way." It's not science, and it doesn't explain anything.

Even total genotypic predictability, however, should give no solace to genetic and darwinian determinism. That's because if everything were pre-determined, then there is no 'natural selection': there's no random mutation screened competitively by the environment, because winners and losers and their genotypes were always already in the works waiting to happen.  In that case, 'evolution' really holds only its original sense of the world: an 'unrolling' of what's already there.  Even before its conception, the future rabbit was doomed to be caught by this particular future fox, in the same sense that we would not say the apple 'evolved' onto Newton's head, but that it just followed the law of gravity.  If so, everything including the misleading (under our working assumption of determinism) evidence of chance (genetic drift, mutation) is simply predetermined!  Unfortunately, perfect determinism cannot be tested or proved, so we're not any further along in our understanding.

The role of 'chance'
The point of science is to understand causation and that is essentially the same as making predictions of the future from observations in the present.  However, even if perfect determinism were the case, things are not predictable in the usual informal sense of deterministic world-views: since we can't have perfect information, only the artifacts of measurement and sampling make a perfectly deterministic world seem non-determined.  We have to make assumptions about when, where, how, and how much our measurements were off--about the distribution and probabilities of what would only appear to be chance.  And those probabilities are themselves determined in some way, if you think about it.

We know from such things as chaos theory among others, that even tiny measurement errors can lead predictions to be completely wrong in a way that's largely unpredictable and untestable (this is often referred to as the 'butterfly effect').  From A to Z we face measurement errors and make assumptions about phenomena that lead us to treat them as driven by chance even if they were entirely pre-determined.

We thus seem to face an inevitable degree of unpredictability even if we're in a wholly deterministic cosmos.  And many physicists argue that the cosmos in fact is at its most fundamental level irreducibly probabilistic (this, for example, in quantum mechanics).  In turn, this means that things cannot be predicted more precisely than by probability--and the conditions under which we can actually know what that probability is, or how it limits the accuracy to which we can aspire, are very limited.

And what if life and its working-out really do have a fundamentally probabilistic component--that is, selection, mutation, mate choice, and genetic interactions with environments?  In that case, at present we generally have scant knowledge of what such probabilities actually are or even how to identify them.  In other words, whether life is or is not wholly deterministic, it at least inescapably appears to be probabilistic. And that means that individuals' natures cannot just be read off a DNA sequencer.

Either, neither, both!

There is some strangeness here.  Think about this:

Natural selection can in principle be
either deterministic, because it moves traits in some direction over time
or probabilistic, because its direction inherently wavers over time
or both deterministic and probabilistic, because a trait can change in a given ‘direction’ over time yet
    not in a straight projection,
or neither deterministic and probabilistic, because a trait doesn’t change in a straight line nor does it just
    change purely randomly.  

As we'll discuss in the next po, one can always see after-the-fact then-to-now evolutionary change and argue that it was somehow determined and inevitable, in a force-like way.  Alternatively, one can argue that it's at least largely luck where things have got to where they are today.  Or, one can say it was deterministic at any given time but that the driving context changed over time.

The very same either-or-both-neither characteristics apply to the causal relationships between an organism's genotype and its traits.  Not all bearers of a given genotype have the same trait value, and a given environment doesn't seem always to have the same effect. Most genetic variants empirically, at as current studies are designed, have small or unstable effects relative to each other; the exceptions are often lethal.  That in turn means that natural selection cannot be a deterministic screen--it can't 'see' each individual's genotype.

We as yet have nothing close to a precise general theory for how deterministic these factors could be, or if not deterministic what causes them to be probabilistic, or if probabilistic what the determines the probability values.  Even in the latter case, we're usually stuck with very limited sampling and observation, not strong theory.

One deep but simple truth might be that DNA is basically an inert molecule that has no function until it interacts with (a physicist might say is 'measured by') its environmental context, however that itself is determined.  Genes and environment are not as simply separable as their separate names indicate.

We should take what we know about what we don't know seriously and make more of an effort to understand it.  Ignorance the the proper driver of science, but only if we acknowledge it.

In the end
Dogma is not helpful here.  Even to acknowledge that natural selection is a mix of determinism and probabilism is a rather empty tautology.  What sort of 'mix'?  Are there general laws that can tell us?  Causation, determinism, probability, uncertainty: are in themselves elusive concepts.

Most scientists acknowledge these facts when they’re pointed out, even if grudgingly when a vested point of view is threatened by that acknowledgment. While the facts we know may require that we gain a deeper insight into what determinism and probabilism actually may mean here, declarations of what is the primary mover are simply not very helpful.

All of this should temper the fervor with which stances are taken on this subject. Opening the door of complexity threatens ideological stances, in science and more acutely when applied to human life and society, but it's the only way to try to gain a deeper understanding, even if once the door is open you can't control the uncertainties that come pouring in.

The drug that reinforces staunch positions are those clear-cut conditions that we do know of, and this goes back to Mendel himself.  But even strongly causal individual alleles, even purportedly lethal alleles, demonstrably vary in their effects and sometimes have essentially no effects.  This is very clear from lots of data, including in humans where it is not just a politically-correct dream-world.  Neither environmentalists nor genomic determinists should shut the door on the nature of biological and evolutionary causation, including the important role of unpredictable chance.

Genes in translation

Years ago, Ken was in Italy and feeling cocky about his command of the language.  He intended to order still, not sparkling, water at a restaurant, so he asked for "Acqua non gazzini," so certain that he'd guessed the word-ending correctly that he was surprised when the waiter laughed, and then so did his friends.  It was a minor error (it's gasata, the past participle), his meaning easily discernible and he got his water, but still, he had mutated the word enough for it to be noticeable.

When I was a kid, we had many foreign exchange students at our house.  Among many memorable, novel English constructions and miscomprehensions that happened around our kitchen table was this one:  The young Honduran student staying with us at the time told my father that she was going to see a movie with a friend.  I don't remember what the movie was, but I do remember my dad teasing her, telling her that it was a movie full of sex and violence, and that it was his job to warn her.  When she came home, he asked how she'd like the film.  She said she was perplexed, in fact.  "I saw the sex, but I didn't see any violins."  

The list of interlingual malapropisms is long, and indeed possibilities endless.  Many are humorous, but some so wrong that the meaning is entirely lost in the uttering.  I was prompted to think about this by a review in a recent New Yorker by Adam Gopnik ("Word Magic: How much really gets lost in translation?").  He opens the piece with an Italian malapropism of his own -- he thought he had ordered little wild strawberries for dessert (fragoline) one night, so was surprised when the waiter brought him a plate of green beans (fagiolini). This has become a favorite family story.  

Gopnik's piece is about languages and translation, as he reviews "what may be the weirdest book the twenty first century has so far produced: 'Dictionary of Untranslatables: A Philosophical Lexicon,'" a hefty book originally in French, but now in English, and much altered from the original.  

It is in part an anti-English protest, taking arms against the imperializing spread of our era's, well, lingua franca--which has now been offered in English, so that everyone can understand it.  The book's presupposition is that there are significant, namable, untranslatable differences between tongues so that, say, "history" in English, histoire in French, and Geshichte in German have very different boundaries that we need to grasp if we are to understand the texts in which the words occur.  

Indeed, histoire in French has an added meaning that 'history' in English doesn't have, including 'history' in the English sense, and 'story'.  Gopnik unleashes a wonderful word that describes words that have multiple meanings in translation -- the Greek word logos, e.g., is a fine example of 'polysemy', with its twenty-three different meanings in English, including 'the word'.  Of course, it's not just in translation that we find polysemy.  Many words serve multiple purposes in any single language too -- in English we've got the many meanings of 'duck' or 'chuck' or 'bust', e.g.

Gopnik is not a fan of the dictionary editor's idea that some words are not translatable, have meaning  so specific in a given language that much is lost in translation.  Contrary to the American linguist Benjamin Whorf, who, early in the 20th century, devised the theory of linguistic relativism, the idea that the language we speak constrains and shapes our thoughts, Gopnik believes that anything can be said in any language, even if it takes more than a single word.

Genes are translated too
Ok, now let's translate all this to genetics -- of course you knew that's where I was headed. Genetics, too, is building its own hefty dictionaries of gene function.  Genes, too, have malapropisms that are noticeable but don't do any damage, just as they have mutations that completely change the meaning.  Many genes have multiple functions, or are pleiotropic, the genetic equivalent of polysemy.  And, as with words, a gene's function is determined by context -- the type of cell it's being expressed in, or what else is going on at the time, developmentally or in response to environmental influences or in neighboring cells or tissues.

But, less prosaically I think, an analogous difference of view to that between language relativists -- or determinists -- and language non-determinists is telling, the dictionaries of genetic function being interpreted in different ways.  Language determines and constrains what we think, Whorf believed, but this view is dated, and most linguists now would argue for a greatly tempered relativism (despite the fact that Jorge Luis Borges once said that he loved writing in English because there's no word for 'wistful' in Spanish), saying that language may have some influence on how and what we think, but isn't overly deterministic.

Similarly, while genes 'for' traits or diseases are still reported all the time, surely the number of human geneticists who would admit to looking for 'the' gene for their favorite trait has diminished, most acknowledging that things are more complex than even they had thought in previous decades.

But, word that there's no gene for wistfulness hasn't spread far and wide outside the field, so we've got political scientists and economists and psychologists and anthropologists and epidemiologists looking for genes to explain their favorite traits.  Linguistic determinism is an arcane theory, with little impact or potential impact on society.  Malapropisms of genetic determinism are another story.

And what about the meaning of the words we use in the field?  There is no consistent definition of 'gene', even in the technical sense.  Just like 'evolution' -- or 'logos' or 'histoire' -- different people utter the word with different things in mind.  Often, probing will show that they weren't very clear even in their own minds about the boundaries and range of the meaning.  Is 'gene' a metaphor for a trait, a protein-coding region of DNA, a protein-coding region including its flanking regulatory DNA, a functional unit visible to natural selection, a single nucleotide that may or may not have a function (and what is a 'function'?).  When we discuss areas that matter, in which there are strong disagreements, semantics can be an important part of the issues at stake.  Often the meaning of the word affects study design, choice of what to study, and explanation of what is found.

It's enough to make one feel wistful for simpler times.

Who, me? I don't believe in single-gene causation! (or do I?). Part IV. Do we need the probabilistic hypothesis?

In the earlier posts in this series on the nature of genetic causation we showed that, while people would routinely say that they don't 'believe' in genetic determinism or single-gene causation, there are clear instances (that everyone recognizes) in which specific genetic variants do seem to have essentially deterministic effects relative to some outcome -- cystic fibrosis is one example, but there are many others.  The data aren't perfect -- there is always measurement noise and other errors, and prediction is rarely perfect -- but in this situation if an individual has a specific single-locus genotype s/he is essentially destined to have the trait it codes for.  Going back to Mendel's famous peas, there is a wealth of animal and plant data for such essentially perfect predictive genotypes.

Yet even there, there are subtleties, as we discussed in earlier installments.  Much of the time, the identified genotype at the 'causal' gene does not predict the outcome with anything resembling certainty.  There is usually evidence for factors that explain this other than lab errors, including variants in other genes, environmental exposures, and so on.

Since most traits seem to be complex far beyond these single-gene subtleties, and are affected by many genes (not to mention environmental exposures), genetics has moved to multilocus causal approaches, of which the most autopilot-like approaches are the omics ones like GWAS (and expression profiling, mirobiomics and many others).  These approaches rest on exhaustively enumerative technology and statistical (inherently and explicitly probabilistic) analysis to find associations that might, upon experimental follow-up, turn out to have a mechanistic causal basis (otherwise, the idea is that the association is a statistical fluke or reflects some unmeasured but correlated variable).  In this context, everyone sneers at the very idea of genetic determinism.  But might that attitude be premature, even for complex traits?

Last week we tried to explain how probabilistic causation comes into the picture. Essentially, each tested variable--each spot along the genome--is screened for variants that are more common in cases than in controls (or in individuals with larger rather than smaller values of some quantitatively measured trait; taller, higher blood pressure, etc.).  But the association is rarely complete: there are cases who don't have the test variant, and unaffected controls who do.  That is a strange kind of 'causation' unless, like quantum mechanics, we invoke fundamental probabilism--but what would be the evidence for that and what does it mean?

The probability hypothesis in genetic causation
As the astronomer Pierre Laplace is famously quoted as saying to Napoleon when asked why his astronomical theory didn't invoke God's causal hand, "Sire, I had no need of that hypothesis." Likewise, we should ask whether, how, or where we have need for the probability hypothesis in genetic causation.  Is probabilism actually the best explanation for what we observe?  If so, how do we find it?

The conclusion is that a genetic variant that passes a statistical significance tests (a deeply problematic notion in itself) has some truly causative effect but one that is in some way probabilistic rather than deterministic.  We assign a 'risk' of the outcome to the presence of the risk factor.  Sample survey methods, which is what the studies are, essentially assume probabilistic outcomes, so of course the results look that way.  But how can a risk factor be probabilistic? We know of some essentially random processes, like mutation, but how do genetic alleles act probabilistically?  One issue is that statistical analysis is fundamentally about repeatable observations....but what does that mean?

In our previous installment, we likened views that seem to be held about multi-genic causation that result from such studies to a conceptual extension of single-gene causal views, but now transformed into single-genometype causation.  But there are serious problems because genomewide, every individual we sample (or who has ever lived, or ever will live) is genomically unique!  And each person is either a case or a control, or has a specific blood pressure level, or whatever.  In that sense, in our actual data, we do not have probabilistic outcomes!  We have one, and only one, specific outcome for each individual.  Yet we don't want to confess to believing in genetic (or genomic) determinism, so we look at one site at a time and assign probabilistic risks to its variants, as if they were acting alone.

Last time we tried to explain serious problems we see with treating each site in the genome (as in, say, GWAS analysis) separately and then somehow trying to sum up the site-specific effects, each treated probabilistically, to get some personalized risk estimate.  We routinely blithely overlook the fundamental issues of the important, usually predominant, effects of environmental factors (most of which are unknown or not accurately estimable).  But, forgetting that huge issue for the moment, surprisingly, maybe deterministic notions of causation are not so far off the mark, after all.

The case for determinism 
The problem is that since in survey sample (epidemiological) research, everyone's genotype is unique we're forced to use probability sampling methods that decompose the genome into separate units, treating it as a sack of loose variants that we can sort through and whose effects we can measure individually.  But to answer the basic questions about genomic epistemology, we do not need to rely so fundamentally on such approaches. Unlike epidemiology, in which each individual is genomically unique, we actually, even routinely do have observation on essentially unrestricted numbers of replications of essentially the exact same genometype!  And that evidence shows a very high level of genetic predictive power, at least relative to given environments.  We have thousands upon thousands of inbred plants and animals, and they have a very interesting tale to tell.

A week or so ago we posted about the surprising level of trait variation due to environmental variation among inbred animals.  But in a standardized environment most inbred strains have characteristic traits, that are manifest uniformly or at least in far higher than few-percent frequency.  This is true of simple inbred strains, or of strains based on selection by investigators for some trait, followed by inbreeding.  (We also have some suggestive but much less informative data on human twins, but for which we only have two observations of each pair's unique genotype, and these are confounded by notoriously problematic environmental issues).

Inbred strains provide close to true replications of genotype-phenotype observations.  There is always going to be some variation, but the fact that inbred strains can be very well characterized by their traits (in a standardized  environment) reveals what is essentially a very high level of genetic--or genometypic--determinism!  So, is the sneering at genetic determinism misplaced?

Informative exceptions
A trait that, even in a standard environment in inbred animals, only appears in a fraction of the animals is particularly interesting in this light.  Such a trait might perhaps aptly be described in probabilistic terms.   We have repeated observations, but only some positive outcomes.  If the environment is really being held constant, then such traits provide good reason to try to investigate what it is that adds the probabilistic variation.  Here, in fact, we actually have a focused research question, to which we might actually get epistemically credible answers, unlike the often total lack of such controllable focus in omic-scale analysis. 

So, when the Jackson Labs description of the strain says the mice 'tend to become hypertensive' or  'tend to lose hearing by 1 year of age', that seems hardly environmental and we can ask what it might mean. Where might such outcome variation come from?  Somatic mutation and the continued influx of germline mutation over generations of the 'same' strain in the lab may account for at least some of it.  Unfortunately, to test directly for somatic mutational effects, one would have to clone many cells from each animal, and that would introduce other sources of variation. But there's a way around this, at least in part:  we can compare the fractions of the variable outcome found in my lab's C57 mice to the fraction of that outcome in the C57s in your lab; this gives you observations on the inbred germline several generations apart, during which some mutations may have accumulated.  If the percentages are similar, they may represent proper and truly (stochastic) probabilities, and we could try to see why and where those effects arise.  If they differ, and your environments really are similar, then mutation would be a suspect.   Such effects could be evaluated by simulation and some direct experimental testing.  It may not be easy, but at least we can isolate the phenomenon, which is not the case in natural populations. 

Not an easy rescue from GWAS and similar approaches!
So, Aha!, you say, inbreeding shows us that genotypes do predict traits, and now all we have to do is use these mice to dissect the causal genes.  Won't that in a sense rescue GWAS approaches experimentally?  Unfortunately, not so!

Interestingly, and wholly consistent with what we've been saying is that if you intercross just two inbred strains, each with its own unvarying alleles so that at most all you have is two different alleles at any spot in the genome (in those sites where the two strains differ), you immediately regenerate life-like complexity:  each animal differs, the intercross population has a typical distribution of trait values (e.g., a normal distribution).  And if you want to identify the genotypic reasons for each animal's trait, you no longer have clonal replicates but must resort back to statistical mapping methods involving the same kinds of false replicability assumptions, and you get the same kinds of complex probabilistic effects, that we see in genomewide association studies in natural populations.  You don't find a neat set of simple additive effects that account for things.

In a sense, despite the simple and replicable genometypic determinism of the parental strains, each intercross offspring has its own unique genometype and trait. This means that the integrated complexity is built into the genometype and not as a rule dissectable into simple causative sites (though as in humans, the occasional strong effect may be found).

Yet, if you took and inbred (repeatedly cloned) each animal in this now-variable intercross population, you would recapture, for each new lineage, its specific fixed and highly predictable phenotype!  And among these animals essentially inbred from the varying intercross animals, with their individually unique genometypes, you would find the 'normal' distribution of trait values.  Each animal's genometype would be highly predictive and determinitive, but among the set you would have the typical distribution of trait values in the intercross population.

That is, a natural population can be viewed as a collection of highly deterministic genometypes with trait effects distributed as a function of the population's reproductive history.

This strong evidence for genometypic determinism might seem to rekindle dreams of personalized genometype prediction, and rescue omics approaches, but it doesn't do that at all.  That's because in itself even a 2-way intercross mapping does not yield simple (or even much simplified) genomic answers for the individuals'  specific trait variants.

Another kind of data is revealing in this light -- transgenic animals: the same transgene typically has different effects in different inbred host strains (e.g., C57, DBA, FVB, ... mice), with the strain-specific effect largely fixed within each strain.  The obvious reason is unique sets of variants in each host strain's background genometype.

Is it time to think differently and stop riding the same horse?  There's no reason here to think anything will make it easy, but there are perhaps nuggets that can be seized and lead to informative approaches and new conceptual approaches. Perhaps we simply have to force ourselves to stop thinking in terms of a genome as a sackful of statistical units, but instead as an integrated whole.

At least, inbreeding reveals both the highly predictive, and hence plausibly determinative, power of genometypes, and provides enormous opportunity for clever research designs to explore genomic epistemology without having to rely on what are the basically counter-factual assumptions of statistical survey analysis.  We can try to work towards an improved actual theory of genetic causation.

But we can't dream of miracles.  Environments could vary, even in the lab as they do in life, and in this sense all bets are off, a topic about which we've posted before.  There seems zero hope of exhaustive DNA enumeration and enumeration of all non-DNA factors, especially in human populations.

At least, things to think about
We hope we are not just luxuriating in philosophical daydreaming here.  There is a lot at stake in the way we think about causation in nature, both intellectually and practically in terms of public resources, public health, disease prediction, and so on. It's incumbent upon the field to wrestle with these questions seriously and not dismiss them because they are inconvenient to think about.

In some ways, the issues are similar to those involved in quantum physics and statistical mechanics, where probabilities are used and their nature often assumed and not questioned, the net result being what is needed rather than an enumeration of the state of each specific factor (e.g., pressure, in the case of the ideal gas law).  The human equivalent, if there is one, might be not to worry about whether causation is ultimately probabilistic or deterministic, nor to have to estimate each element's probabilistic nature specifically, but to deal with the population--with public health--rather than personalized predictions.  But this goes against our sense of individuality and the research sales pitch of personalized genomic medicine.

At least, people should be aware of the issues when they make the kinds of rather hyperbolic claims to the public about the miracles genomics is claiming to deliver, based on statistical survey epistemology.  Genomes may be highly predictive and determinitive, at least in specific environments, but the lack of repeatable observations in natural populations raises serious questions about the meaning or usefulness of assuming genetic determinism.

Ism-itis!

People adhering to a particular faith, say X, are often called Xists or those who believe in Xism.  Sometimes this is just a descriptor, but often it's used as a criticism, or worse.  That's what happens when pole-headed right-wingers call President Obama a 'socialist' (something the pole-heads apparently know nothing about).

Sistine Chapel; Wikimedia Commons

This kind of ism/ist description applies in cultural combat but also in science.  Thus people can be reductionist, frequentist, or falsificationist and so on, in the words of their detractors -- so these aren't compliments.  People actually would not use the word about themselves if they feel that others use the term in a derogatory way.

Creationists (and they really are ists in the sense of holding tightly to a specific ideology or creed!) often denigrate people who understand the real world as 'evolutionists',as it is not clear exactly what these terms mean, which can be important.  It is a castigating characterization when used by creationists, and it may imply that one is an ideolog about, rather than an explorer of, the subject of evolution.

We often refer critically on MT, to evolutionary or Darwinian or genetic 'determinists' or 'determinism.'   However, our meaning is important to understand.  At a seminar in our department a couple of years ago, when we questioned the nature of genetic determinism being invoked in some darwinian adaptive Just-So story-telling, someone stated that he was a determinist--"and isn't everyone?" Well, in science the answer is clearly yes, and no.

Here's where language gets in the way.  If determinism means that Nature is causal and that every effect must have some cause, then most scientists would plead guilty to the charge.  To invoke effects without cause is to be mystic, and certainly that has nothing to do with science.  That doesn't mean that we have identified or understand the causation of effects under discussion, and there is where the legitimate issues lie.

Darwin, Museum of Natural
History, London;
Wikimedia Commons

Things are less transparent when it comes to 'selectionism'.  If one makes the Darwinian assumption that whatever is here had to have got here by adaptive natural selection, then it is perfectly legitimate not to be a selectionist.  Much in evolutionary reconstruction is of this type, and it often includes behavioral evolution or even morphology, where traits themselves can be hard even to define. If the trait must (by assumption) be the result of specific adaptive natural selection, then our task is to identify that selection.  But we can always find some such reason, since the function today can be equated to having been the advantageous function in the past.  This is entirely circular, and it's not science!

This doesn't mean that eyes or birds' wings or hominid locomotory apparatus got here 'by chance', as creationists still falsely often suggest, but it can mean that some functional elements ended up in our genomes by chance, if their initial harmfulness or helpfulness was slight compared to the populations they were in (in genetic terms, they got into the genome by 'drift').  Duplicate genes that have no harmful effect but provide redundancy that can subsequently be used for new function constitute one of many examples.

Selection must build on what's there, however it got there.  Such chance-installed elements don't suddenly produce wings out of reptilian forelimbs, because complex traits involve too many changes.  But the elements themselves need not have got here by selection, since most will have slight effect.  Likewise, truly harmful things are eliminated by not being viable, and that is a form of selection, but not Darwinian or adaptive selection, since the defunct forms weren't really competing with anybody or anything.  They just didn't work.

The bottom line here is that determinism depends on the degree to which (1) truly probabilistic cause exists, and/or (2) one believes that a specific cause under consideration perfectly predicts a specific outcome, and/or (3) the cause acts alone but only has predictive power if very accurately measured, and we can't get such quality measures.  If a causal effect is truly probabilistic, it does not in the usual sense 'determine' the outcome.  And if the cause is but one of many contributors, and hence has weak predictive power, or does so if inaccurately measurable, then arguing for 'determinism' stretches the truth and merits criticism.  It doesn't mean there is no cause, but in these instances we cannot reliably or accurately predict the outcome from observing the cause. 

Likewise for adaptive selectionism or 'Darwinism'.  If selection is weak, sporadic, erratic, or distributed over many different contributing factors, or if there is no selection but only drift, or if there is selection but we have no serious way to argue what its mechanism was, then the adaptationist argument stretches the truth and merits criticism.  Knowing the genes involved in a trait doesn't predict their change from on generation to the next and, indeed, different genotypes can generate the same phenotype, so that we cannot infer the cause from the result.  Again, this doesn't mean there is no cause, but it does mean that selectionism is over-stated.

What we argue when criticizing what we think are excessive claims of genetic determinism or selectionism is that the assertion being made does not bear scrutiny in these above senses.  We're not arguing for mystical causation or effects that are not 'determined' by physical causes.  We may be arguing that we have little idea or way of knowing what the cause(s) was or were, or that the assumption of a kind of causation can be made self-fulfilling rather than really scientifically testable.  It often seems to be true that based on methods and criteria we use today, some of these causal situations simply cannot in principle be worked out beyond some very imperfect level of precision.  Things too probabilistic, or too weak to be understood from the kinds of samples we can actually collect, are simply not accurately predictable from observing putative causes, and that also means we have inadequate ways of even identifying the causes.

We do seem to live in an orderly causal world.  There are 'laws' of Nature, even if the word is a human one that doesn't imply a law-giver.  Whether causation can be truly probabilistic, and whether there may be causal aspects whose very existence humans have not learned to detect or characterize, we have no idea.  Nor does anybody else.  There are wild theories of multiple universes to get around pure probabilistic causes, and things like dark matter and energy to get around some of what we observe in ordinary matter.  Who knows what else someone may some day discover.  Given this, we believe that more circumspection is in order about causal claims in the life sciences.  In part this is because science has practical implications for society, can be used towards evil or harmful ends (even if unintentionally), and costs resources that could be used for other things, if we had a less lobbying-based or ideological social environment in terms of making such decisions.

To argue that someone is a 'determinist' is not to label them with a slur as if they should instead by a mystic or crystal ball reader.  It is to argue that assertions should be tempered, and we should take more seriously the things that are clearly inadequately known but could be quite fundamental.  To be a 'Darwinian' or 'adaptationist' can mean not just that one recognizes the clear truth of evolution as a fact, and that survival requires success by definition, but can refer to someone who goes beyond that, to and assumes what is to be shown, and that certainly is not good science.  One wants to have an interpretive framework, without which science would be difficult if not impossible, but the framework needs to be tested rather than assumed.

Assuming a framework--being an 'ist'--may be good for hustling attention or grants, but not for a more serious--if avowedly less complete--understanding of things that we now have or than the ists of this world would lead you to believe.

Free will in the genomics age: does it have any meaning?

In the March 10 edition of the BBC Radio 4 program In Our Time, the three guests discussed the subject of free will.  Does it exist?  Could it exist in the age of science? Or is it just a mistaken notion that is a hangover from religion, related to what is needed in order for people to be responsible for their own actions, and hence where they end up in Eternity?

This discussion was by philosophers, not neuroscientists, but the neuroscience and general-science perspective is there.  The discussion is quite interesting.  The idea of free will arises because we so much feel that we have it, that we make decisions.

The issue for us relates to the concept of determinism.  If the universe is completely Newtonian, that is, follows perfect laws at all scales of observation, then everything is related to and in that sense predictable by anything.  At the time of the Big Bang, what you are going to have for dinner was, in principle, predictable.  That would have fundamental consequences for evolution, since in a purely deterministic universe there is no real competitive factor among rabbits and foxes: the slowest rabbit was fore-ordained to be dinner for the fast fox.  Random variation screened by unpredictable experience is not what's going on, despite the Modern Synthesis claims to the contrary!

Of course this is all nonsense, because to see that everything is predictable you probably would have to be outside the universe to observe it, but the  idea of a totally deterministic universe is that it's entirely of itself--no outside agent that could meddle, or even observe it.

Anyway, if this is the world we're in, then nobody is responsible for their actions in the moral sense, not even Stalin or Hitler or Ghaddafi, or Mother Theresa for that matter.  We are determined from conception by our genes, and our environment.  Our neurons wire up during life, in totally predictable ways (predictable in principle, that is, if one knew where every molecule and every neural cell was at every instant, etc.), and so the thoughts we think are just the result of that wiring--not in any sense freely thought by us, if thoughts really are just signals flying around among neurons.

As the discussion in the BBC program points out, even if randomness exists, we aren't morally at free will, because we're the combination of pure physical determinism, plus chance events that we don't control.  Thus the usual appeal to quantum mechanics probability doesn't change the story of determinism vs moral responsibility.

But are even 'random' events like mutation really random?  If they are not, but are just determined in ways we can't understand, the world returns to deterministic laws-of-Nature status.
But what is a 'chance' event?  Is it one with no cause?  Or is there some kind of cause that is probabilistic--clearly something we do not understand?  If, for example, random mutations really follow some laws of probability, then determinism just takes a slightly different form.  Free will remains in the realm of the non-material, and hence mystic and non-existent illusion.

So, if our thousands of genes controlling the behavior of billions of cells, in environments with many chance factors, are just working out the physical forces, there is no such thing as free will, no matter how it feels to us.  If there is true probability (neurons wire to some extent just by chance, truly), then there may indeed be something that would genuinely approach free will: it would not be predictable, even by probability distributions (because the latter would not be pure chance, but a different kind of cause).  In terms we understand, at least, pure chance is an effect without a cause!

More likely, what we're learning by all our omics technologies is essentially that things appear so random, and there is so much of it, that we can never, even in principle collect enough data to predict whether you'll have this or that flavor ice cream today, or whether you'll have chicken or pasta on your next overseas flight.  Even if the appearance of randomness in brains, like that in tossed coins, is really just an illusion of randomness.

It is thus hard to escape that no matter how it looks, all that seems to be free will is illusion, not true free will.  And it's dispiriting to feel that so much of life is an illusion (a view that Darwin is supposed to have expressed, though we don't remember seeing the quote - if you know it, let us know).  But if free will is an illusion, mistaken appearance of causeless effects, then for the very same reasons, so is natural selection.  And that is food for thought, for people as well as the happy, not really just lucky, fox.