Is genetics still metaphysical? Part VI. What might lead to a transformative insight in biology today--if we need one

It's easy to complain about the state of the world, in this case, of the life sciences, and much harder to provide the Big New Insights one argues might be due.  Senioritis makes it even easier: when my career in genetics began, not very much was known.  Genes figuratively had 2 alleles, with measurable rates of recurrence by mutation.  Genetically tractable traits were caused by the proteins in these genes; quantitative traits were too complex to be considered seriously to be due to individual genes, so were tacitly assumed to be the additive result of an essentially infinite number of them.

How many genes there were was essentially unknowable, but using identified proteins as a gauge, widely thought to be around 100,000. The 'modern evolutionary synthesis' solved the problem, conceptually, by treating these largely metaphorical causal items as largely equivalent, if distinct, entities whose identities were essentially unknowable.  That is, at least, we didn't have to think about them as specific entities, only their collective actions.  Mendelian causal genes, evolving by natural selection was, even if metaphorical or even in a serious way metaphysical, a highly viable worldview in which to operate.  A whole science enterprise grew around this worldview.  But things have changed.

Over the course of my career, we've learned a lot about these metaphysical units.  Whether or not they are now more physical than metaphysical is the question I've tried to address in this series of posts, and I think there's not an easy answer--but what we have, or should have, understood is that they are not units!  If we have to have a word for them, perhaps it should be interactants.  But even that is misleading because the referents are not in fact unitary.  For example, many if not  most 'genes' are only active in context-dependent circumstances, are multiply spliced, may be post-trascriptionally edited, are chemically modified, and have function only when combined with other units (e.g., don't code for discretely functioning proteins), etc.

Because interaction is largely a trans phenomenon--between factors here and there, rather than just everything here, the current gene concept, and the panselectionistic view in which every trait has an adaptive purpose, whether tacit or explicit, is a serious or even fundamental impediment to a more synthetic understanding. I feel it's worth piling on at this point, and adding that the current science is also pan-statistical in ways that in my view are just as damaging.  The reason, to me, is that these methods are almost entirely generic, based on internal comparison among samples, using subjective decision-criteria (e.g., p-values) rather than testing data against a serious-level theory.

If this be so, then perhaps if the gene-centered view of life, or even the gene concept itself as life's fundamental 'atomic' unit, needs to be abandoned as a crude if once important approximation to the nature of life. I have no brilliant ideas, but will try here to present the sorts of known facts that might stimulate some original thinker's synthesizing insight--or, alternatively, might lead us to believe that no such thing is even needed, because we already understand the relevant nature of life: that as an evolutionary product it is inherently not 'regular' the way physics and chemistry are.  But if our understanding is already correct, then our public promises of precision medicine are culpably misleading slogans.

In part V of this series I mentioned several examples of deep science insight, that seemed to have shared at least one thing in common:  they were based on a synthesis that unified many seemingly disparate facts.  We have many facts confronting us.  How would or might we try to think differently about them?  One way might be to ask the following questions: What if biological causation is about difference, not replication?  What if 'the gene' is misleading, and we were to view life in terms of interactions rather than genes-as-things?  How would that change our view?

Here are some well-established facts that might be relevant to a new, synthetic rather than particulate view of life:

1. Evolution works by difference, not replication Since Newton or perhaps back to the Greek geometers, what we now call 'science' largely was about understanding the regularities of existence.  What became known as 'laws' of Nature were, initially for theological reasons, assumed to be the basis of existence.  The same conditions led to the same outcomes anywhere.  Two colliding billiard balls here on Earth or in any other galaxy, would react in identical ways (yes, I know, that one can never have exactly the same conditions--or billiard balls--but the idea is that the parts of the cosmos were exchangeable.)  But one aspect of life is that it is an evolved chemical phenomenon whose evolution occurred because elements were different rather than exchangeable.  Evolution and hence life, is about interactions or context-specific relative effects (e.g., genetic drift, natural selection). 

2.  Life is a phenomenon of nested (cladistic) tree-like relationships Life is not about separated, independent entities, but about entities that from the biosphere down (at least) to individual organisms are made of sets of variations of higher-level components.  Observation at one level, at least from cells up to organs to systems to individuals, populations, species and ecosystems, are reflections of the nested level(s) the observational level contains. 

3.  Much genetic variation works before birth or on a population level Change may arise by genetic mutation, but function is about interactions, and success--which in life means reproduction--depends on the nature of the interactions at all levels.  That is, Darwinian competition among individuals of different species is only one, and perhaps one of the weakest, kinds of such interaction.  Embryonic development is a much more direct, and stronger arena for filtering interactions, than competition (natural selection) among adults for limited resources.  In a similar way, some biological and even genetic factors work only in a population way (bees and ants are an obvious instance, as are bacterial microfilm and the life cycles of sponges or slime molds). 

4.  Homeostasis is one of the fundamental and essential ways that organisms interact Homeostasis as an obvious example of a trans phenomenon.  It's complexly trans because not only do gene-expression combinations change, but they are induced to change by extra-cellular and even extra-organismal factors both intra and inter-species.  The idea of a balance or stasis, as with organized and orchestrated combinatorial reaction surely cannot be read of in cis.  We have known about interactions and reactions and so on, so this is not to invoke some vague Gaia notions, but to point out the deep level of interactions, and these depend on many factors that themselves vary, etc. 

5. Environments include non-living factors as well as social/interaction ones No gene is an island, even if we could identify what a 'gene' was, and indeed that no gene stands alone is partly why we can't.  Environments are like the celestial spheres: from each point of view everything else is the 'environment', including the rest of a cell, organ, system, organism, population, ecosystem.  In humans and many other species, we must include behavioral or social kinds of interactions as 'environment'.  There is no absolute reference frame in life any more than in the cosmos.  Things may appear linear from one point of view, but not another.  The 'causal' effects of a protein code (a classical 'gene') depend on its context--and vice versa. 

6. The complexity of factors often implies weak or equivalent causation--and that's evolutionarily fundamental. Factors or 'forces' that are too strong on their own--that is, that appear as individually identifiable 'units'--are often lethal to evolutionary survival.  Most outcomes we (or evolution) care about are causally complex, and they are always simultaneously multiple: a species isn't adapting to just one selective factor at a time, for example.  Polygenic causation (using the term loosely to refer to complex multi-factoral causation) is the rule.  These facts mean that individually identified factors usually have weak effects and/or that there are alternative ways to achieve the same end, within or among individuals.  Selection, even of the classical kind, must be typically weak relative to any given involved factor. 

7. The definition of traits is often subjective and affects their 'cause' Who decides what 'obesity', 'intelligence', or 'diabetes' is?  In general, we might say that 'Nature' decides what is a 'trait', but in practice it is often we, via our language and our scientific framework, who try to divide up the living world into discrete categories and hence to search for discrete causal factors.  It is no surprise that what we find is rather arbitrary, and gives the impression of biological causation as packaged into separate items rather than being fundamentally about a 'fabric' of interactions.  But the shoehorn is often a major instrument in our causal explanations. 

8. The 'quantum mechanics' effect: interaction affects the interactors In many aspects of life, obviously but not exclusively applied to humans, when scientists ask a question or publicize a result, it affects the population in question.  This is much like the measurement effect in quantum mechanics.  Studying something affects it in ways relevant to the causal landscape we are studying.  Even in non-human life, the 'studying' of rabbits by foxes, or of forests by sunlight, affects what is being studied.  This is another way of pointing out the pervasive centrality of interaction.  Just like political polls, the science 'news' in our media, affect our behavior and it is almost impossible to measure the breadth and impact of this phenomenon.

All of these phenomena can be shoe-horned into the 'gene' concept or a gene-centered view of life or of biomedical 'precision'.  But it's forced: each case has to be treated differently, by statistical tests rather than a rigorous theory, and with all sorts of exceptions, involving things like those listed here, that have to be given post hoc explanations (if any). In this sense, the gene concept is outmoded and an overly particulate and atomized view of a phenomenon--life--whose basic nature is that it is not so particularized.

Take all of these facts, and many others like them, and try to view them as a whole, and as a whole that, nonetheless can evolve.  Yesterday's post on how I make doggerel was intended to suggest a similar kind of mental exercise.  There can be wholes, and they can evolve, but they do it as wholes. If there is a new synthesis to be found, my own hunch it would be in these sorts of thoughts.  As with the examples I discussed a few days ago (plate techtonics, evolution itself, and relativity), there was a wealth of facts that were not secret or special, and were well-known. But they hadn't been put together until someone thinking hard about them, who was also smart and lucky, managed it. Whether we have this in the offing for biology, or whether we even need it, is what I've tried to write about in this series of posts.

Of course, one shouldn't romanticize scientific 'revolutions'.  As I've also tried to say, these sorts of facts, which are ones I happen to have thought of to list, do not in any way prove that there is a grand new synthesis out there waiting to be discovered. It is perfectly plausible that this kind of ad hoc, chaotic view of life is what life is like.  But if that's the case, we should shed the particulate, gene-centered view we have and openly acknowledge the ad hoc, complex, fundamentally trans nature of life--and, therefore, of what we can promise in terms of health miracles.

Is genetics still metaphysical? Part V. Examples of conditions that lead to transformative insights

A commenter on this series asked what I thought that "a theory of biology should (realistically) aspire to predict?" The series (part 1 here) has discussed aspects of life sciences in which we don't currently seem to have the kind of unifying underlying theory found in other physical sciences. I'm not convinced that many people even recognize the problem.

I couched the issues in the context of asking whether the 'gene' concept was metaphysical or was more demonstrably or rigorously concrete.  I don't think it is concrete, and I do think many areas of the life sciences are based on internal generic statistical or sampling comparison of one sort of data against another (e.g., genetic variants found in cases vs controls in a search for genetic causes of disease), rather than comparing data against some prior specific theory of causation other than vacuously true assertions like 'genes may contribute to risk of disease'.  I don't think there's an obvious current answer to my view that we need a better theory of biology, nor of course that I have that answer.   

I did suggest in this series that perhaps we should not expect biology to have the same kind of theory found in physics, because our current understanding doesn't (or at least shouldn't) lead us to expect the same kind of cause-effect replicability.  Evolution--which was one of the sort of basic revolutionary insights in the history of science, and is about life, specifically asserts that life got the way it is by not being replicable (e.g., in one process, by natural selection among different--non-replicate--individuals).  But that's also a very vanilla comment.

I'll try to answer the commenter's question in this and the next post.  I'll do it in a kind of 'meta' or very generic way, through the device of presenting examples of the kind of knowledge landscape that has stimulated new, deeply synthesizing insight in various areas of science.

1.  Relativity
History generally credits Galileo for the first modern understanding that some aspects of motion appear differently from different points of view.  A classic case was of a ship gliding into the port of Genoa: if someone inside the ship dropped a ball it would land at his feet, just as it would for someone on land.  But someone on land watching the sailor through a window would see the ball move not just down but also along an angled path toward the port, the hypotenuse of a right triangle, which is longer than the straight-down distance.  But if the two observations of the same event were quantitatively different, which was 'true'?  Eventually, Einstein extended this question using images such as trains and railroad stations: a passenger who switched on two lightbulbs, one each at opposite ends of a train, would see both flashes at the same time.  But a person at a station the train was passing through would see the rearmost flash before the frontward one.  So what does this say about simultaneity?

These and many other examples showed that, unlike Isaac Newton's view of space and time as existing in an absolute sense, they depend on one's point of view, in the sense that if you adjust for that, all observers will see the same laws of Nature at work.  Einstein was working in the Swiss patent office and at the time there were problems inventors were trying to solve in keeping coordinated time--this affected European railroads, but also telecommunication, marine transport and so on. Thinking synthetically about various aspects of the problem led Einstein later to show that a similar answer applied to acceleration and a fundamentally different, viewpoint-dependent, understanding of gravity as curvature in space and time itself, a deeply powerfully deeper understanding of the inherent structure of the universe.  A relativisitic viewpoint helped account for the nature and speed of light, aspects of both motion and momentum, of electromagnetism, the relationship between matter and energy, the composition of 'space', the nature of gravity, of time and space as a unified matrix of existence, the dynamics of the cosmos, and so on, all essentially in one go.

The mathematics is very complex (and beyond my understanding!).   But the idea itself was mainly based on rather simple observations (or thought experiments), and did not require extensive data or exotically remote theory, though it has been shown to fit very diverse phenomena better than former non-relativisitc views, and are required for aspects of modern life, as well as our wish to understand the cosmos and our place in it.  That's how we should think of a unifying synthesis. 

The insight that led to relativity as a modern concept, and that there is no one 'true' viewpoint ('reference frame'), is a logically simple one, but that united many different well-known facts and observations that had not been accounted for by the same underlying aspect of Nature.

2.  Geology and Plate Techtonics (Continental Drift)
Physics is very precise from a mathematical point of view, but transformative synthesis in human thinking does not require that sort of precision.  Two evolutionary examples will show this, and that principles or 'laws' of Nature can take various forms.  

The prevailing western view until the last couple of centuries, even among scientists, was that the cosmos had a point Creation, basically in its present form, a few thousand years ago.  But the age of exploration occasioned by better seagoing technology and a spirit of global investigation, found oddities, such as sea shells at high elevations, and fossils.  The orderly geographical nature of coral atolls, Pacific island chains, volcanic and earthquake-prone regions was discovered.  Remnants of very different climates than present ones in some locations were found.  Similarly looking biological species (and fossils) were found in disjoint parts of the world, such as South Africa, South America, and eventually Antarctica.  These were given various local, ad hoc one-off explanations.  There were hints in previous work, but an influential author was Alfred Wegener who wrote (e.g., from 1912--see Wikipedia: Alfred Wegener) about the global map, showing evidence of continental drift, the continents being remnants of a separating jigsaw puzzle, as shown in the first image here; the second shows additional evidence of what were strange similarities in distantly separated lands.  This knowledge had accumulated by the many world collectors and travelers during the Age of Exploration. Better maps showed that continents seemed sometimes to be 'fitted' to each other like pieces of a jigsaw puzzle.  

Geological ages and continental movement (from Hallam, A Revolution in the Earth Sciences, 1973; see text)

Evidence for the continental jigsaw puzzle (source Wikipedia: Alfred Wegener, see text)

Also, if the world were young and static since some 'creation' event, these individual findings were hard to account for. This complemented ideas by early geologists like Hutton and Lyell around the turn of the 19th century. They noticed that deep time also was consistent with the idea of (pardon the pun) glacially slow observable changes in glaciers, river banks, and coastlines that had been documented since by geologists  Their idea of 'uniformitarianism' was that processes observable today occurred as well during the deep past, meaning that extrapolation was a valid way to make inferences.  Ad hoc isolated and unrelated explanations had generally been offered piecemeal for these sorts of facts.  Similar plants or animals on oceanically separated continents must have gotten there by rafting on detritus from rivers that had been borne to the sea.

Many very different kinds of evidence were then assembled and a profound insight was the result, which we today refer to by terms such as 'plate techtonics' or 'continental drift'.   There are now countless sources for the details, but one that I think is interesting is A Revolution in the Earth Sciences, by A. Hallam, published by Oxford Press in 1973, only a few years after what is basically the modern view had been convincingly accepted.  His account is interesting because we now know so much more that reinforces the idea, but it was as stunning a thought-change as was biological evolution in Darwin's time.  I was a graduate student at the time, and we experienced the Aha! realization that was taking place was that, before our very observational eyes so to speak, diverse facts were being fit under the same synthesizing explanation (even some of our faculty were still teaching old, forced, stable-earth explanations).

Among much else, magnetic orientation of geological formations, including symmetric stripes of magnetic reversals flanking the Mid-Atlantic Trench documented the sea-floor spreading that separated the broken-off continental fragments--the pieces of the jigsaw puzzle.  Mountain height and sea depth patterns gained new explanations on a geologic (and very deep time) scale, because the earth was accepted as being older than biblical accounts).  Atolls and the volcanic ring of fire are accounted for by continental motions.  

This was not a sudden one-factor brilliant finding, but rather the accumulation of centuries of slowly collected global data from the age of sail (corresponding to today's fervor for 'Big Data'?).  A key is that the local facts were not really accounted for by locally specific explanations, but were globally united as instances of the same general, globally underlying processes.  Coastlines, river gorges, mountain building, fossil-site locations, current evidence of very different past climates and so on were brought under the umbrella of one powerful, unifying theory.  It was the recognition of very disparate facts that could be synthesized that led to the general acceptance of the theory.  Indeed, subsequent and extensive global data, continue to this day to make the hypothesis of early advocates like Wegener pay off.

3.  Evolution itself
It is a 100% irrefutable explanation for life's diversity to say that God created all the species on Earth. But that is of no use in understanding the world, especially if we believe, as is quite obvious, that the world and the cosmos more broadly follows regular patterns or 'laws'.  Creationist views of life's diversity, of fossils, and so on, are all post hoc, special explanations for each instance. Each living species can be credited to a separate divine reason or event of creation.  But when world traveling became more common and practicable, many facts and patterns were observed that seemed to make such explanations lame and tautological at best.  For example, fossils resembled crude forms of species present today in the same area.  Groups of similar species are found living in a given region, with clusters of somewhat less similar species elsewhere. The structures of species, such as of vertebrates, or insects, showed similar organization, and one could extend this to deeper if more different patterns in other groups (e.g., that we now would call genera, phyla, and so on).  Basic aspects of inheritance seemed to apply to species, plant and animal alike.  If all species had been, say, on the same Ark, why were similar species so geographically clustered?

It dawned on investigators scanning the Victorian Age's global collections, and in particular Darwin and Wallace, that because offspring resemble their parents, though are not identical to them, and individuals and species have to feed on each other or compete for resources, that those that did better would proliferate more.  If they became isolated, they could diverge in form, and not only that but the traits of each species were suited to its circumstances, even if species fed off each other.  Over time this would also produce different, but related species in a given area.  New species were not seen directly to arise, but precedents from breeders' history showed the effects of selective reproduction, and geologists like Lyell had made biologists aware of the slow but steady nature of geological change.  If one accepted the idea that rather than the short history implied by biblical reading, life on earth instead had been here for a very long time, these otherwise very disparate facts about the nature of life and the reasons for its diversity might have a common 'uniformitarian' explanation--a real scientific explanation in terms of a shared causative process, rather than a series of unrelated creations: the synthesis of a world's worth of very diverse facts made the global pattern of life make causal and explanatory sense, in a way that it had never had before.

Of course the fact of evolution does not directly inform us about genetic causation, which has been the motivating topic of this series of posts.  We'll deal with this in our next post in the series.

Insight comes from facing a problem by synthesis related to pattern recognition
The common feature of these examples of scientific insight is that they involve synthesis derived from pattern recognition. There is a problem to be solved or something to be explained, and multiple facts that may not have seemed related and have been given local, ad hoc, one-off 'explanations'. Often the latter are forced or far-fetched, or 'lazy' (as in Creationism, because it required no understanding of the birds and the beasts). Or because the explanations are not based on any sort of real-world process, they cannot be tested and tempered, and improved.  And, unlike Creationist accounts, scientific accounts can be shown to be wrong, and hence our understanding improved.

In our examples of the conditions in which major scientific insights have occurred, someone or some few, looking at a wealth of disparate facts, or perhaps finding some new fact that is relevant to them, saw through the thicket of 'data', and found meaning.  The more a truly new idea strikes home, in each case, the more facts it incorporates, even facts not considered to be relevant.

Well!  If we don't have diverse, often seemingly disparate facts in genetics then nobody does!  But the situation now seems somewhat different from the above examples: indeed, with the precedents like those above, and several others including historic advances in chemistry, quantum physics, and astronomy, we seem to hasten to generalize, and claim our own synthesizing 'laws'.  But how well are we actually doing, and have we identified the right primary units of causation on which to do the same sort of synthesizing?  Or do we need to?

 I'll do my feeble best to offer some thoughts on this in the final part of this series.