A Systems-Theory Perspective on Cancer

 

Comments on Werner J. Schwemmler’s

Basic Cancer Programs – Genes, Signals, Metabolites –

Unified Holistic Theory of Evolution

 

 

Ben Goertzel

October 23, 2002

 

 

1. Introduction

 

About a month ago, in the midst of some quasi-random web-surfing on biomedical topics, I came across the abstract of a fascinating book by Werner J. Schwemmler, with the whopper title

 

Basic Cancer Programs

Genes, Signals, Metabolites

Unified Holistic Theory of Evolution

 

The book was published in 1998, but it seems not to have been terribly widely distributed, at least not in the US (the author and publisher are German).   I ordered it directly from the publisher, Karger, and found it an extremely stimulating read.  

 

Perhaps the most impressive things about Schwemmler’s work are its breadth of scope, and its gutsiness.   The author is not afraid to put forth big, bold hypotheses – and he has the knowledge and imagination and persistence to back up his seemingly wild proposals with empirical and theoretical details.  This kind of combination of breadth and detail is rare in any scientific discipline, but particularly so in biology, which tends to be relatively non-theoretical in nature. 

 

Schwemmler’s Basic Cancer Programs is a rarity among rarities in the bio literature: a book that presents not one but two original, far-reaching, fascinating, and reasonably precise biological theories.   Put roughly, the two theories are:

 

 

The first of these theories is not new in Basic Cancer Programs, but was presented previously in Schwemmler’s1989 book Symbiogenesis, A Macromechanism of Evolution: Progress Toward a Unified Holistic Theory of Evolution, as well as in numerous prior research papers.  The second was not presented in book form previously to Basic Cancer Programs, though it was discussed in a few earlier technical research papers.  The two theories stand somewhat separately from each other – either one could be falsified or verified independently of the other – but they are closely conceptually related.

 

Before plunging into the thick of Schwemmler’s ideas, I’ll make a few comments about the book itself.   It is a brief book – under 150 pages -- but not an easy one.  The “level” is very uneven.  The topics are of extremely broad interest, and some parts of the book could be read by the educated layperson – but other parts are clearly written by a biologist for other biologists.  I have a moderate knowledge of several branches of biology, but I’m not a biologist, and I found some portions of Basic Cancer Programs rather rough going.  The discussion is generally well structured, and the author continually draws useful links between general principles and specific biological examples.  The numerous diagrams are also helpful.  But the density of biological concepts and terminology is great, and I found myself frequently leafing through the substantial glossary.   Also – and this is probably the key expository flaw -- Schwemmler doesn’t always do all he could to outline his key ideas in advance of presenting the mass of details supporting them.  The discussion often seems to follow Schwemmler’s own train of thought, which is interesting, but is the train of thought of an expert in every biological subdiscipline.

 

These difficulties notwithstanding, however, I found the effort to digest Schwemmler’s ideas very worthwhile, and I would encourage you to make a similar effort.  I do not know if his ideas are correct -- at this stage, no one knows that.  However, they are plausible, interesting hypotheses connecting the organismic, cellular and molecular levels, and having such hypotheses is a valuable thing.   Even if not correct, they serve to stimulate thought in new directions, both theoretically and in terms of the development of experimental and analytical tools. 

 

In this brief essay on Basic Cancer Programs, I have two interconnected goals.  The main goal is purely expository: to present the crux of Schwemmler’s hypotheses about symbiogenesis and cancer in a way that can easily be understood by the educated non-biologist.   But there is also a secondary goal, which is more creative: to explore, in a general way, the means by which his ideas about cancer can be validated or falsified using computational analysis of genomic and proteomic time series data collected using microarrays.   I won’t harp on this topic extensively here, but due to my personal interest in microarray data analysis, it has been on my mind throughout my study of Schwemmler’s work, so I will mention it here and there.  Schwemmler’s theory, combined with modern techniques for analyzing microarray time series data, suggests new approaches to identifying chemical or genetic therapies capable of arresting cancer growth or even causing cancerous cells to revert to healthy cellular form.

 

 

2. Symbiogenesis

 

Schwemmler’s unique take on evolutionary biology is centered on the notion of symbiogenesis: symbiosis between organisms as a key mechanism of the creation of new biological form. 

 

Classic neo-Darwinist evolutionary theory posits that new forms evolve via mutation, crossover and selection.  Complex-systems-oriented, self-organizationist evolutionary theory acknowledges the importance of neo-Darwinist mechanisms, but emphasizes the role of complex physical dynamics in the creation of biological structures.  Some theorists – such as A. Lima de Faria in Evolution Without Selection (1990) -- have taken the self-organizationist approach to an extreme, arguing that self-organization is the cause of all evolution, with neo-Darwinist “differential reproduction based on fitness” playing a minimal or nonexistent role.  Stephen Jay Gould was a vocal advocate of a more moderate approach, in which self-organization and traditional mutation/crossover/selection share the leading role.  This moderate approach was reviewed elegantly by Augros and Stanciu in their book A New Biology (1985); and many related ideas were discussed in my 1993 book The Evolving Mind.

 

Symbiogenesis introduces a new ingredient to the “selection vs. self-organization” issue.  The notion is best-known in an origin-of-life context, due to Lynn Margulis’s (1991) provocative hypothesis that life originated via the fusion of two symbiotic proto-organisms.  Schwemmler accepts Margulis’s hypothesis and takes it further, by analyzing aspects of contemporary cellular behavior as a regression to the proto-cellular symbiotic state.

 

He introduces the term “endocytobiology” as a blanket term encompassing both symbiosis and parasitism – or in his words, “to include both intracellular symbionts (endocytosymbionts) and intracellular parasites (endocytoparasites), taking into account the fact that endocytobiosis is a fluid transition between symbiosis and parasitism.”   This term was introduced by Schwemmler in 1979 based upon the already-existing term “endocytobiosis.”

 

Schwemmler reviews evidence for the role of symbiogenesis in the evolution of species.   Most provocatively, he emphasizes the role of inter-cellular and inter-species gene transfer.  As he says, in endocytobiological cell theory,

 

the genomes of host and symbiont cells can be intracellularly synchronized, then partly mutually exchanged (horizontal gene transfer), and finally they can fuse (nuclear and cellular fusion)….   In addition to the intracellular horizontal gene transfer, the transfer of genes between spatially separated organisms and even organisms of different species may occur.   For example, plants synthesize leghemoglobin, which is homologous to the hemoglobin of animals.  From this fact an interspecific gene transfer can be deduced.

 

The skeptic might point out that there is no really hard evidence for symbiogenesis as the cause of any particular examples of speciation.  On the other hand, it must be admitted that there isn’t much hard evidence for any detailed theories of species creation, a fact that creationists have used to argue against natural selection as an explanation of the origin of species.

 

Schwemmler proposes symbiogenesis as an explanation for cases in which vastly different organisms display similar forms; e.g. he discusses an insect called the lanter carrier that shares many structural similarities with the crocodile, and notes that

 

The endocytobiological model at least provides a general explanation for the structure and function of the crocodile mimicry of the lantern carrier.

 

I don’t find this entirely convincing, in that non-symbiogenetic self-organizational morphogenetic dynamics may also explain this sort of mimicry.  Maybe the structures and functions shared by the lantern carrier and the crocodile are “attractors” of a certain general biological dynamic shared by both species (see, Goertzel, 1993 for a relevant discussion).

 

On the other hand, Schwemmler’s discussion of symbiogenesis in the evolution of the leafhopper – an insect he spent much of his career studying – is significantly more convincing.   He postulates that bacteria have transferred some of their genes to the gene pool of the leafhopper, and that in doing so, they have help the leafhopper acquire its ability to digest various plant substances:

 

Apparently, the symbionts mediate between the host as the consumer and its nutriment or prey as the producer by compensating for chemical and physiological differences through decomposition processes.

 

 

Only in this way were leafhoppers able to adapt to changing plant nutrition such as mosses, ferns, and seed plants.   The phylogenetic function of endocytobionts seems to be to mediate chemically and physiologically between the lower physiochemical type of the host insect and the higher type of its food sources. (p. 74)

 

 

Finally – as if this isn’t bold enough -- Schwemmler ties in symbiogenesis with a variant of Ernst Haeckel’s controversial notion that “ontogeny recapitulates phylogeny” (see Gould, 1985 for a modern treatment of Haeckel’s ideas).   The growth of cells from stem cells into fully differentiated and socialized cells, he argues, bears great resemblance to the primordial origin of life out of nonlife.

 

In an argument of great significance for his theory of cancer, he elaborates this hypothesis in terms of the two chemical processes of fermentation and respiration.  During early embryogenesis, he points out, fermentation rather than respiration dominates, in both insects and mammals.  A little later on, when cell differentiation occurs, respiratory processes gradually take over; and ultimately, once a cell is irreversibly differentiated, the total decline of fermentative metabolism results the loss of the capability for cell division.   Cell development may thus be viewed as a combination of fermentative division processes without differentiation, and respiratory differentiating processes without division.   And this, he points out, is a plausible retelling of what happened long ago when life first emerged from the primordial soup.  A fermentative proto-cell symbiotically merged with a respiratory proto-cell, producing the first cell.   The symbiogenetic origin of life, he suggests, is important for understanding the dynamics of cell development within organisms existent today.

 

3. Fermentation, Respiration, and the Basic Programs of Cell Development

 

 

Following standard practice, Schwemmler divides the cell development process into several phases.  He characterizes these phases as five “basic cellular programs”:

 

  1. Division
  2. Growth
  3. Differentiation
  4. Function
  5. Apoptosis (cell death)

 

Each of these programs, he points out, contains components pertaining to three different cellular subsystems: the cell nucleus, plasm and membrane. 

 

His analysis of the specific programs involved here is tied in with his thinking on the relationship between fermentation and respiration in primordial proto-cells.   He constructs the following chart, showing the relationship between fermentation and respiration at different stages of cell development:

 

 

No fermentation

Apoptosis (cell death)

 

Function: irreversibly differentiated cells

Aerobic fermentation (G2)

 

Growth, reversible differentiation (late embryonic cells for instance)

 

Anaerobic fermentation (G1)

Division (totipotent stem cells; early embroyonic cells)

 

 

 

No respiration (A)

Anaerobic respiration (A1)

Aerobic respiration (A2)

 

Fig. 40, p. 83

 

 

The question he poses is: Why are only four of the nine cells in this table occupied?   The general answer is: Due to the dynamics induced by the activator and suppressor genes associated with the pertinent programs.  But this general answer doesn’t go far enough.  To really understand cancer or any other significant cellular phenomenon, one needs to know the particulars of the genetic networks involved with these stages of the cell cycle, in the nucleus, cytoplasm and cell wall.

 

Never at a loss for hypotheses, Schwemmler has something to say here as well.  In Figure 45 (which is large and complex and not reproduced here), he give a

 

 hypothetical circuit diagram of signals, genes and metabolites for the regulation of the five basic cellular programs and their possible derailment into the regressive carcinogenesis program.

 

I’ll turn to the carcinogenesis aspect in the following section; for now let’s stick with “simple” healthy-cell regulation.   What Figure 45 does, in this regard, is to outline the particular genes and proteins, and the categories of genes and proteins, involved in key roles in the five stages of the cell cycle, including their impacts on each of the three major parts of the cell.  

 

I don’t believe that Figure 45 is complete, or correct in all its details.   In all probability, Schwemmler himself would now have modifications to make, given that 4 years have passed since Basic Cancer Programs was published, and a lot of biology has happened since then.  I do believe, however, that this is precisely the kind of diagram we ought to be drawing, if we want to understand cell development, cancer, or just about any other complex biological phenomenon.  

 

One exercise that would be extremely instructive would be a detailed comparison of Schwemmler’s diagram with the metabolic and signal pathways given in the Kegg pathways database (http://www.genome.ad.jp/kegg/kegg5.html).  I have carried out this exercise to a limited extent already, in an informal and unsystematic way.  Not surprisingly, there is a lot of overlap; but there are also a lot of pathways in Kegg whose relationships to Schwemmler’s schematic are not clear (at least to me).  I would love to see a variant of Figure 45 based on integrating Kegg’s pathway information, and using the general theoretic notions of Schwemmler and others to fill in the gaps.   Such a thing does not seem to exist in the biology literature, a fact presumably attributable to the magnitude of the task, and the tendency of researchers in this area to focus on individual genes, molecules and pathways rather than on the overall wiring diagram.

 

 

4. Carcinogenesis as Regression

 

 

But what does Schwemmler actually say about cancer?  In good systems theory style, he sees cancer not as an isolated and particular phenomenon, but as an aspect of the behavior of complex systems with complex histories.

 

In a nutshell, he views carcinogenesis as a kind of reversal of the five basic cellular programs.  The following chart (inspired by, but not directly drawn from, Basic Cancer Programs) summarizes his five stages of cancer development:

 

 

 

Stages of Carcinogenesis

Relation to the Cell Maturation Process

1

Initiation (formation of the cancer “stem cell”)

 

2

Promotion (activation of the cell cycle; deterioriation of intercellular communcation)

De-differentiation; reactivation of cell division

3

Progression (loss of brakes on cell division, and of responsiveness to external control signals)

Increasing fermentative activity, more complete de-differentiation

4

Invasion (metastasis: membrane permeability to sugars; cell migration)

Almost complete de-differentiation; respiration halted

5

Death

 

 

 

Schwemmler makes several major observations about this multi-phase process of cancer development.  

 

Firstly, he cites research noting that cancerous cells often begin making use of fermentative rather than respiratory processes to gather energy.  This is very significant, in Schwemmler’s perspective, because it signifies a move back toward embryonic processes. 

 

Secondly, cancer cells become de-differentiated as cancer progresses.  Healthy cells are very diverse, both in their structures and functions, and in their communications with other cells.  On the other hand, all metastasized cancer cells are fairly similar.   Cancer progression is in a way a process of re-homogenization.  All cell types began similar, and cancer returns them to a similar state.  And it is no coincidence, he posits, that the return to fermentation occurs along with dedifferentiation.  Both of these things signify a return to the primal state.

 

Following this up, he posits that cancer cells’ partial emulation of the embryonic condition may be part of the key to their survival.   The body is programmed to protect embryonic cells, and this programming may cause it to erroneously protect cancer cells as well.

 

 

The great functional resemblance of cancer and embryonic cells has far-reaching consequences, leading to the organism’s hazardous acceptance of and care for the tumor as if it were an embryo, indeed to the organism’s protecting the tumor at the expense of its own health.

 

 

When they turn on fermentation, cancer cells seem to re-activate functions which in various ways resemble the basic functions of embryonic cells.  (p. 91)

 

 

All this doesn’t tell us what causes cancer.   But, if correct, it tells us a lot about the carcinogenetic process. 

 

Regarding the cause of this process, Schwemmler also has a theory, interlocking nicely with his ideas on symbiogenesis and the origin of life.   He begins with the relatively commonplace observation that cancerous cells possess multiple mutations which must cooperate in order to launch the carcinogenetic process.  As he says,

 

The genomes of some tumor cells contain mutations at up to 12 different sites….  These mutations mainly affect signal receptors and antigen patterns on the cell surface, the relation of fermentation to respiration and the gene activation by transcription factors in the nucleus.  (p. 94)

 

His particular twist on this idea has to do with the relation of viral genes to mutated cancerous genes.  As he notes,

 

In retroviruses, genes have been identified that encode for proteins which are homologous to different elements of various signal chains….  In comparison to the cell genes, the virus genes show typical alterations.  These structural gene alterations, however, are mostly rather small.   If such viral genes become expressed by a cell, their transcription products can cause disturbances in the cell and possibly its transformation into a tumor “stem” cell as well.   Due to their carcinogenic nature, these viral genes are classified as oncogenes.  Thus, viral oncogenes are nothing but cellular activator genes, so-called proto-oncogenes, which have been mutated in the virus….

 

Besides these RNA viruses there are also DNA viruses that can influence tumor formation.  These viruses activate the cellular DNA replication system by blocking intrinsic division suppressor genes of the cells.   Such cellular genes are, therefore, called anti-oncogenes or tumor suppressor genes.  (p. 94)

 

In short, virii may interact with cells in such a way that their genes become expressed in the cell.  These virii may simultaneously express mutated versions of many different genes – a much more plausible idea than the hypothesis that such a large set of simultaneous mutations with emergently interacting effects should simply emerge by chance.

 

And this ties in with the fermentation/respiration theme.  In cancer cells, the genes involved in mitochondrial respiration are mutated and functionally disturbed.  According to Schwemmler’s endocytobiological cell theory, this can be described as a partial “dissolution” of the original symbiotic fusion that led to the origin of the primordial cell (in Schwemmler’s terms, a reversal of the endocytobiosis between the anaerobic urkaryotic host and its formerly aerobic eubacterial symbionts, which later became the mitochondria). 

 

To a degree, then, it seems tumor formation can be regarded as a return to the original state -- meaning that carcinogenesis represents a phylogenetic regression.  Like the early proto-cells in the primordial soup, a cancer cell is characterized by the unlimited ability to divide, by a primitive glycolysis, and by the inability to cooperate with other cells within a multicellular organism.  The parallel between cancer cells, embryonic cells and primordial proto-cells is not complete, but it is deep and interesting, and seems likely to stimulate very interesting experimental and theoretical research.

 


5.  Testing Schwemmler’s Hypotheses

 

As should be clear from the foregoing summary, Basic Cancer Programs presents a number of interrelated hypotheses, some of which are more easily testable than others. 

 

The role of symbiogenesis in the evolution of species falls into the “very hard to test” category.  Certainly it can be verified or refuted, with reasonable confidence, via an appropriate combination of experiments at various biological levels.  But there seems no single experimental approach likely to provide a definitive judgment on the idea.  At least, none occurs to this writer at the moment, and Schwemmler does not provide any. 

 

One interesting possible research path, peripherally related to Schwemmler’s symbiogenetic ideas, lies in the artificial life domain.  It would be very interesting to run genetic-algorithm-based simulations of speciation, and observe whether the introduction of symbiogenetic phenomena significantly increases the frequency of speciation, the effectiveness of new species, the complexity of emergent ecological webs, and so forth.

 

On the other hand, Schwemmler’s hypothesis of carcinogenesis as regression is significantly more manageable in the short run.  The data is not available now to verify or refute it, but, reasonable mechanisms for collecting this data would seem to exist, in the form of gene and protein microarrays.  The regulatory networks to which Schwemmler refers are largely inferable from time series of gene and protein expression levels.  What one needs is a collection of gene and protein expression time series data sets referring to various cells going through various stages of both embryogenesis and carcinogenesis.  From such data, one could determine whether it is actually true that the carcinogenetic process involves the same series of “regulatory programs” as the embryogenetic process, though in reverse.  One could create a variant of Schwemmler’s Figure 45, based on the mining of empirical data, rather than on intuitive information integration. 

 

Of course, given truly complete and perfect data on genetic and proteomic dynamics, one wouldn’t need theories like Schwemmler’s at all.  One could just crunch the data, see what the pathways are, and start thinking about the best pharmacological approaches to halting carcinogenesis.  But the reality is that microarrays and other related technologies don’t give us complete and perfect data, but rather incomplete and noisy data.  And, given this reality, the Schwemmler’s ideas and other speculative theories have a rather large role to play.  When the data is clear, the data must guide the theory (and potentially confirm or refute the theory).  But when the data is ambiguous, the theory must guide the data interpretation.

 

In short, Schwemmler’s theory of carcinogenesis is too large to be verified or refuted by any single experiment, but it would be susceptible to a coordinated program of data collection and analysis carried out over a period of several years.  And the technology is currently available to carry out this program.  We have a theory that is not only broad in scope and precise in implications, but intensely experimentally explorable in the short term.

 

6. The Role of AI-Driven, Information-Integrative Data Analysis

 

Now I turn, briefly, to my own personal interest in Schwemmler’s ideas.   I am not a biologist, but rather (among other things) a bioinformaticist -- a computer scientist involved in creating software for analyzing biological data.  I’m specifically concerned with “microarray data” – data gathered from microarrays regarding the expression levels of genes and proteins in various cells under various conditions (see Kohane et al, 2002, for an in-depth discussion of genetic microarrays).   As a bioinformaticist, I find Schwemmler’s ideas extremely interesting, because of the guidance they may provide my software in analyzing very difficult datasets. 

 

Together with my colleagues at Biomind LLC, I’m currently developing several microarray-oriented software products, one of which, the Biomind Genetic Network Analyzer, addresses the “regulatory network inference” problem.  That is, this product takes in a microarray time series data set, and attempts to answer the question: What “regulatory program” (s) are observable in this dataset during a certain time interval?   The recent research literature contains a number of approaches to this problem, none of which is entirely satisfactory, but many of which show promise (see Wessels et al, 1999, for a review of some of these techniques).  Our approach is based on the sophisticated techniques embodied in the “Novamente AI Engine” general-purpose AI software system (Goertzel and Pennachin, in preparation), and we’re quite excited about it. 

 

But no matter how sophisticated your algorithms, to get really confident results from regulatory network inference software is very difficult at this stage.  In most cells, there will be a lot of gene activity at any given time (even during “stationary phase”), and identifying the major coordinated networks of activity is far from simple, especially given the noise level involved in contemporary microarray technology. 

 

One strategy we’re using to work around these problems is the integration of diverse background knowledge into the data analysis process.   There is a huge amount of information in biological and chemical databases, pertinent to the inference of genetic networks from microarray data.  At present the human mind is the only tool capable of integrating this information into the data analysis process.  But we believe this problem of broad-based data integration is amenable to AI techniques.  And we believe that ideas like Schwemmler’s may play a role here.  Just as one may integrate facts from databases into the data analysis process, so one may also integrate hypothetical ideas.  

 

As a very simple example of the integration of speculative hypotheses into the data analysis process, consider the notion of an inhibitory relationship between fermentative and respiratory processes.  Existing databases contain annotations regarding which genes and proteins are involved in fermentation versus respiration.  The theoretical hypothesis that fermentative and respiratory processes are mutually inhibitory may be tested via looking at gene and protein expression time series datasets.  But it may also be used to guide the analysis of such datasets.  If one knows that a given protein is involved in aerobic fermentation, and another protein is involved in aerobic respiration, this means one should assign a relatively low a priori probability to the simultaneous expression of these two proteins.  If one is using a learning method such as Bayesian statistics, which relies on a priori probabilities, this theoretical supposition may lead to different data analysis results than one would obtain without it.  (My team’s work doesn’t involve traditional Bayesian statistics, but it uses probability theory in a related way, which does make use of a priori information.) 

 

The value of incorporating hypothesis into data analysis is easy to see in the simple example of the inhibitory relationship between fermentation and respiration, but a similar point applies to more complex hypotheses about particular metabolic and signal pathways.  One needs a data analysis method with a complex knowledge representation in order to incorporate something like Schwemmler’s Figure 45 as “prior probabilistic information,” but this is not an insurmountable problem, particularly not in the context of an integrative AI framework like Novamente.

 

 

 

7. Toward Novel Cancer Therapies

 

 

Finally, what does Schwemmler have to suggest regarding novel cancer therapies, based on his ideas about cancer development? 

 

He really doesn’t give any specific suggestions for new pharmacological solutions.  Rather, he suggests a new strategy for approaching the problem.  “Map the genetic networks involved” isn’t exactly a novel idea, but he has a lot of specific ideas about how these genetic networks work, and this is very refreshing, given the focus of most of the oncogenomics literature on very small sets of genes and proteins.

 

In addition to giving a lot of suggestions likely to be helpful for genetic network mapping, Schwemmler’s theories also suggest some concrete pleasingly indirect directions for cancer therapy research.   For instance: Perhaps, rather than trying to directly suppress cancerous cell behavior, or kill cancerous cells, we should actively seek to induce cancer cells to display healthy behaviors.  If Schwemmler’s hypotheses about regulatory program interactions are correct, then cell differentiation and carcinogenetic reproduction consist largely of sets of mutually inhibitory programs.  The question then becomes, how do we induce cell differentiation, so as to reverse the regressive process?

 

And this – as with nearly all other potential applications of Schwemmler’s ideas -- comes down to detailed knowledge of the genetic networks in question.  What are the genes whose expression triggers activation of the “cell differentiation” program?   From within this gene set, are there some genes that can be introduced without leading to adverse effects?   Are there some that are particularly easy to introduce into cells via well-understood “gene therapy” transport mechanisms?

 

Schwemmler’s insights, at this stage, are far from providing a cure for cancer.  What they do give, however, is a new way of looking at the problem.   This new perspective fits in naturally with new data collection and analysis methods – microarrays and associated time series analysis algorithms – and also with the increasing trend toward systems-biology thinking in general.

 

 

8. The Power of Systems Biology

 

Biology is full of fascinating empirical discoveries, deep conceptual insights, and detailed theoretical analyses of particular phenomena.  But my mind, trained in mathematics, physics and computing, persists in asking “where are the big biological theories?” – the theories that are broad in scope, precise in implication, and reasonably well empirically substantiated? 

 

There are a few.   The theory of evolution by natural selection is one example; the genetic code another.  In the neuroscience domain, Edelman’s theory of neuronal group selection attempts to stake out this ground.  But these examples are few and far between.  Of course, it would be fatuous to try to impose the model of the physical sciences on biology.  Biological systems are different than other physical or chemical systems – “messier,” and in some senses more complex – and the different nature of biological knowledge reflects this.  There will be no Newton’s Laws or Schrodinger Equation or valence bond theory for biology; living systems aren’t like that.  But recognition of the different nature of biology shouldn’t stop us from pushing as far as possible in the direction of a more deeply theoretical biology.  The recent increase in literature on “systems biology” represents, in part, an emerging desire on the part of a large section of the biology community to move in the direction of integrative conceptual understanding.

 

Schwemmler, in Symbiogenesis and now Basic Cancer Programs, has given us a new addition to the short list of “big biological theories.”  And for this we should be grateful.   His ideas on carcinogenesis and genetic networks should be valuable guidance over the next few years, as the mapping of the genome is followed up by the mapping of genetic and proteomic networks.  His ideas on symbiogenesis and speciation may be more difficult to empirically investigate in the near future, but should be seriously considered by all theorists and experimentalists working on the fundamentals of evolution. 

 

 

 

References