Toolbox murders: putting genes in their epigenetic and ecological contexts
Griffiths and Stotz’s Genetics and Philosophy: An Introduction offers a very good overview of scientific and philosophical issues raised by present-day genetics. Examining, in particular, the questions of how a “gene” should be defined and what a gene does from a causal point of view, the authors explore the different domains of the life sciences in which genetics has come to play a decisive role, from Mendelian genetics to molecular genetics, behavioural genetics, and evolution. In this review, I highlight what I consider as the two main theses of the book, namely: (i) genes are better conceived as tools; (ii) genes become causes only in a context. I situate these two theses in the wider perspective of developmental systems theory. This leads me to emphasize that Griffiths and Stotz reflect very well an on going process in genetics, which I call the “epigenetization” of genetics, i.e., the growing interest in the complex processes by which gene activation is regulated. I then make a factual objection, which is that Griffiths and Stotz have almost entirely neglected the perspective of ecological developmental biology, and more precisely recent work on developmental symbioses, and I suggest that this omission is unfortunate in so far as an examination of developmental symbioses would have considerably strengthened Griffiths and Stotz’s own conclusions.
KeywordsGene Genetics Epigenetics Development Information Symbiosis
Genetics and Philosophy: An Introduction (later on GP) is a very useful and timely book. As an introduction to genetics and the philosophy of genetics, it is a very accessible book, though it explores recent and complex data. It offers a well-informed and stimulating overview of current genetics in the so-called “post-genomic era” (Eisenberg et al. 2000) and of recent philosophical debates about genes (What is a gene? How to identify one gene? What should we do of the different and non-overlapping definitions of the term “gene” that co-exist in the biological literature? Are genes causally “special”? What is the relation between genes and the “nature/nurture” debate?) Presenting these fundamental issues in a clearly written and short book (250 pages) was a key challenge. This challenge has been met with remarkable success by Griffiths and Stotz.
Beyond, GP also constitutes an excellent introduction to philosophy of biology in general, as it combines an impressive expertise in current molecular biology with an important contribution to major philosophical issues (in particular about biological causation). Within philosophy of biology, molecular biology has been much less extensively explored than evolutionary biology, a surprising situation given the wonderful transformations of molecular biology in the last 60 years (see, for example, Morange 1998; Sarkar 2005). This book will certainly count as one important exception to this relative neglect for molecular biology (along, of course, with a few others, for instance Rosenberg 1985, 2006; Schaffner 1993; Sarkar 1998, 2005; Beurton et al. 2000; Keller 2000; Moss 2003; Burian 2005; Weber 2005; Craver 2007; for an overview, see Darden and Tabery 2009).
Finally, GP is also exemplary in that it makes welcome excursions into the history of biology—for instance concerning the debate between the Mendelians and the biometricians about heredity (Chapter 2), or the construction of molecular genetics and the transition from an understanding of biological “specificity” based on stereochemistry to one based on information (Chapter 3).
The overall structure of the book is as follows. After an introductory chapter, Chapter 2 examines the birth of “Mendelian” genetics at the beginning of the twentieth century. Following Falk (1984), this chapter distinguishes between the instrumental gene (defined by its role in genetic analysis) and the hypothetical material gene (the postulated material unit of heredity), and shows that Mendelian genetics focused on the gene understood as an instrumental unit, indispensable to account for observed correlations between the phenotypes of parents and offspring. Chapter 3 tells the story of the quest for the material gene, with the definition of a “gene” as a fragment of DNA that specifies the linear order of its products (RNAs, proteins) (this is the “classical molecular gene”). This chapter shows that the Mendelian gene has not been reduced to the new molecular gene; the two concepts of the gene simply play two different roles, and these two concepts still co-exist today. Chapter 3 also recalls how the notion of biological specificity, initially related to the idea of stereochemistry, became an informational notion in the second half of the 1950s, mainly under the influence of Crick (1958). Chapter 4 starts the analysis of present-day, “post-genomic” molecular biology. It shows that informational specificity, far from being located in DNA alone, is actually distributed among many different factors, including regulatory RNAs and proteins, which leads the authors to reject reductionist approaches and to defend a view that they call, after Burian (2004), “molecular epigenesis.” Chapter 5 examines the way environmental factors influence genome expression and contribute to informational specificity in Crick’s sense; this analysis leads the authors to question the nature/nurture dichotomy, and to suggest the notion of a “developmental niche”. Chapter 6 examines the ideas of genetic information and genetic program; though the authors confirm some of their previous critiques against genetic information, they express more sympathy with the idea that the “genetic code” can be understood as one key evolutionary innovation by which informational specificity is faithfully transmitted from one generation to the next. Chapter 7 builds on previous work by Griffiths and Tabery to examine how genetics is used in explanations of behaviour, and in particular human behaviour.
The Introduction holds the promise of offering “new lessons for philosophy of biological science”, based on recent transformations of genetics and molecular biology. This raises the issue of the extent to which GP offers new perspectives in philosophy of biology and in philosophy of science. In her recent review of a book by Evelyn Fox Keller, Stotz (2012) has set a very high bar for qualifying as an innovating work in philosophy. Using the metaphor of Agatha Christie’s Murder on the Orient Express, Stotz suggests that Keller’s book does not deliver the fatal blow to the nature/nurture dichotomy, but is just one of the many killers. If one applies such high standards, it seems to me that GP will not qualify as entirely innovating. Yet two answers can be given. First, GP presents itself as an introductory work (and is highly successful as such), and obviously novelty is not the principal objective of an introductory work. Second, and more importantly, this book does bring new ideas, or rather new and important arguments and data in favour of some philosophical claims already expressed in the past, but which are considerably strengthened by these new arguments and data (especially in Chapters 4, 5 and 6 with regard to biological causality).
It would be impossible, naturally, to cover in the present review all the different aspects examined in this very rich book. I focus on what I see as some of the most important views defended by Griffiths and Stotz, with a particular emphasis on philosophical issues. The “toolbox murders” of my title is not so much a reference to Tobe Hooper, the highly acclaimed director of The Texas Chain Saw Massacre and Poltergeist, as an emphasis on what is, in my view, the main lesson of GP, namely that genes are best conceived astools. This idea is applied differently to two very important meanings of the word “gene”. First, the Mendelian gene is a tool in the sense that it is an instrument used by geneticists to account for and predict the results of the interbreeding between two organisms (analysed in Sect. “The failure of reductionism and the survival of the Mendelian gene” below), and, importantly, this gene concept still exists in some branches of today’s biology. Second, the molecular gene is best understood not as a causally autonomous and deterministic factor in development, but rather as a tool for the cell (analysed in Sect. “The “epigenetization” of genetics” below). So the take-home message is that genes should be seen not as much as architects, builders, or even tinkerers than as tools—though very important ones, of course.
I consider that the most convincing contribution of GP lies in the analysis of how current molecular genetics has been transformed due to the adoption of an increasingly “epigenetic” point of view (what I call “the epigenetization of genetics”). Indeed, recent years have witnessed the accumulation of data showing the implication of regulatory RNAs and regulatory proteins in the most fundamental genetic processes, including transcription and translation. In light of this transformation, Griffiths and Stotz explore with remarkable insight the philosophical issue of causality (examined in Sect. “Re-thinking developmental causality” below). The implication of many epigenetic mechanisms in “genetic” processes strengthens the idea of “distributed causality”, that is, the idea that the causality of development is shared among many and interacting factors, without the possibility to single out genetic factors as causally special.
In Sect. “The missing final blow: developmental symbioses”, I explain why I think the authors should have paid more attention to the field of “ecological developmental biology”, and in particular to the many instances of developmental symbioses, and I show why those examples would have been very helpful to strengthen several theses defended in this book.
The multiple identities of the gene
Building on previous work by many biologists, historians and philosophers (among whom Falk 1984, 1986, 2000; Portin 1993; Burian 1995, 2005; Keller 2000; Moss 2003; and also Griffiths and Neumann-Held 1999; Griffiths and Stotz 2006), Griffiths and Stotz show that the concept of a “gene” has several, non overlapping meanings. Historian of science Rafael Falk has shown that, as early as the Mendelians, the gene had two identities: first, an instrumentalidentity, according to which the gene is defined by its role in genetic analysis; second, an hypothetical material identity, according to which the gene is a physical entity the exact nature of which has to be discovered. Importantly, the success of the instrumental gene in genetic analysis did not depend on the elucidation of its physical nature: presupposing “elements” or “factors”, without any specification about the material nature of these “elements” or “factors”, was enough to make predictions about interbreeding work. Many Mendelian geneticists hoped that this physical nature would soon be determined, but this was a desirable addition to the perspective of genetic analysis, not a necessary condition to make it work.
Moss (2003) contributed to this debate by distinguishing between “genes-P” (for “phenotypes”, “prediction” and “preformation”) and “genes-D” (for “development”). “Genes-P” are the basis used to explain phenotypes in hybridization experiments, so they are close to Falk’s “instrumental genes”. “Genes-D” are defined as template for the making of gene products, in particular proteins, so they are close to the “material genes” of molecular biologists.
The failure of reductionism and the survival of the Mendelian gene
Griffiths and Stotz start with the examination of the “Mendelian gene”, which offers the first illustration of why genes are often better understood as tools. This first sense concerns the domain of classical genetics: for a “Mendelian” biologist, the “gene” is to be understood as an instrument for predicting the result of interbreeding experiments. Faced with the results of numerous interbreeding experiments, and in particular the proportions of re-occurring characters that disappear in the first generation and reappear in the next, Mendelians found that the postulation of the transmission of “factors”, and later “genes”, was simply the best available explanation. Biologists were confident that their “genes” had a good explanatory power, regardless of their material nature, as Morgan famously emphasized in his 1933 Nobel lecture. In addition, applying the Mendelian approach made it possible to design new experiments and obtain important new results, as illustrated by the work of Morgan’s group showing that in Drosophila females have two X chromosomes and males have one X and one Y chromosome. Therefore, in the context of Mendelian genetics, even if there were no straightforward physical particles corresponding to genes, genes would still be essential devices for calculation. In Falk’s (1984, 1986) terms, the main identity of genes in Mendelian genetics is their instrumental identity, precisely because their role is to explain and predict observable results of interbreeding experiments, and to suggest new experiments.
Importantly, the subsequent story of gene was an evolution (new definitions were attributed to the notion of a gene), but not a reduction. In particular, it is simply not true, as Griffiths and Stotz argue after many others, that, with the progress made by molecular biology in the 20th century, the “Mendelian gene” has been reduced to the “molecular gene”. There are two reasons for this: first, there is no straightforward matching between the Mendelian gene and the molecular gene; second, the Mendelian gene is still alive and well in some domains of contemporary biology. The first point is illustrated by the clear fact that not all segments of chromosomes that behave as Mendelian alleles count as genes under the new molecular conception (for example, untranscribed regulatory regions not immediately adjacent to the coding sequences they regulate can segregate independently of those coding sequences, and so can function as separate Mendelian alleles, but, under most definitions of a molecular gene, they are not separate molecular genes). It was initially hoped that the molecular gene would be at once the unit of replication, the unit of mutation and the unit of function, but the molecular gene today is not the unit of replication (this role is played by the whole DNA molecule), and it is not the unit of mutation (this role is played by a single DNA nucleotide); it is only the unit of function (namely, it plays the role of producing a gene product, typically an RNA molecule or a protein). The second point is illustrated by the fact that the Mendelian gene still appears in several domains of current biology, particularly quantitative genetics.
In agreement with the quasi-consensus reached in philosophy of biology over the last 40 years (e.g., Hull 1974; Kitcher 1984; Sarkar 1998; Griffiths and Neumann-Held 1999; Gilbert and Sarkar 2000; Burian 2005), Griffiths and Stotz conclude that the gene does not offer a case of successful reductionism. After Sarkar (1992) and Brigandt and Love (2008), they distinguish different kinds of reductionism (methodological reductionism; ontological reductionism, epistemic reductionism). For Griffiths and Stotz, contemporary molecular biology has been “reductionist” methodologically, in the sense that it has been highly successful in offering explanations situated at the low level of molecular interactions, but even this “reductive” methodology often needs to be complemented with a more “integrative” approach, as emphasized by many biologists today (e.g., Noble 2006). Ontologically, there is a consensus on physicalism (biological entities are, in the end, “nothing but” physicochemical entities). From an epistemic point of view, recent molecular biology clearly does not confirm theory reduction, nor does it support reduction understood in terms of modes of analysis, but it does illustrate a form of explanatory reduction, as already suggested by Sarkar (2005), Weber (2005), and Rosenberg (2006). Yet, even this “explanatory reductionism” should be conceived as both a reductive and an integrative strategy, because it is fundamental to integrate molecular details into a unified picture that takes into account all interactions and feedback loops (Gilbert and Sarkar 2000; Craver and Bechtel 2007).
The “epigenetization” of genetics
Chapters 4, 5 and 6 are probably the most stimulating chapters of the book, because they present exciting scientific data accumulated in the last two decades and offer in addition rich philosophical insights, in particular concerning the problem of biological causality and the notion of specificity. Griffiths and Stotz make a strong case for “molecular epigenesis”, the thesis according to which development, even at the molecular level, is always the result of a complex interplay between genes, regulatory mechanisms, and the environment. The term “molecular epigenesis” was used by Burian (2004, p. 59) to describe “the revision of sequence-based information by altering molecular conformations or by action of non informational molecules”, under the influence of the cellular and external environments, and then by Stotz (2006). Of course, this perspective is also in part reminiscent of Waddington, who coined the term “epigenetics” in 1942, and highlighted how genes are activated and regulated in development and also how development is always the product of the interactions of many causal factors (e.g., Waddington 1955). Here, Griffiths and Stotz offer a detailed explanation of the thesis of molecular epigenesis by systematically examining the different ways in which genes are regulated through different and interacting RNA- and protein-based mechanisms, and how these regulatory mechanisms are involved in developmental causality and specificity.
Research done in the current “postgenomic era” has led to several striking results. First, only 1.5 % of the sequenced DNA corresponds to protein-coding genes; the rest includes genes for noncoding RNA, pseudogenes, introns, untranslated regions of mRNA, regulatory DNA sequences, repetitive DNA sequences, and sequences related to mobile genetic elements. Second, many RNAs are non-coding RNAs but exert key functional roles, in particular concerning the regulation of protein production. Third, there are around 20,000–25,000 genes in humans, but many more proteins (possibly between 250,000 and one million), so, contrary to what was thought in the 1960s, the specificity of proteins as gene products cannot be reduced to the specificity of gene sequences.
A first consequence is that the structural gene (the traditional understanding of a gene as a sequence of nucleotides giving rise to a gene product, typically a protein) has tended to become something like a “consensual” gene (Fogle 2000), meaning that the structural gene is often just what biologists decide to define as a gene. “Gene annotation” often rests explicitly on such consensual definitions as, for example, when geneticists decide where a gene starts and ends in the face of the widespread phenomenon of alternative splicing. Burian (2004) has talked about “nominal genes” to refer to structural genes understood this way, that is, as consensually defined sequences, which function as useful devices for communicative purposes within the scientific community and sometimes beyond.
The second consequence is that, in particular because of the frequent phenomenon of gene combination, it seems difficult to maintain the traditional idea that the molecular gene must be at once a structural (a well-delineated and reasonably continuous sequence of nucleotides) and a functional unit (that which makes a gene product, typically a protein). One possible answer to this problem is to accept to step back from structural definitions of the gene, and to offer a purely functional definition of the gene, which is exactly what Griffiths and Stotz do with their “postgenomic gene” concept. According to such a definition, a gene is any sequence or set of sequences that can make a product, no matter how structurally complex they may be. Here the structural unity of the gene often becomes so fragmented and complex that it ceases to be an adequate focus, the only concern being to offer a functional account of what a gene produces. This is the view endorsed by Griffiths and Stotz (see also the “molecular process gene concept” in Griffiths and Neumann-Held 1999), and also by Gerstein et al. (2007).
Two lessons can be drawn from the adoption of this functional definition of the molecular gene: (i) A “gene” is much more than just a sequence of DNA. This leads to the idea of “molecular epigenesis” defended by the authors; (ii) “Genetic causality” involves much more than just DNA sequences, leading to a far-reaching re-conceptualization of biological causality. The first lesson is examined here, while the second is discussed in the next section.
Research done in genetics in the last 20 years has shown that a “gene” is much more than a sequence of DNA, because of the crucial importance of gene activation, combination, and regulation. When cells receive environmental signals (related for example to hormonal or nutritional changes), they change their gene expression via the environment-specific use of regulatory mechanisms. Such regulatory mechanisms of genome expression can activate, select and create sequence information, often in a causally specific way. An example of sequence activation mechanism is the phosphorylation of transcriptional regulators. An example of sequence selection is alternative splicing, the phenomenon by which different versions of mature mRNA transcripts can result from the cutting and joining of different combinations of exons. In alternative splicing, a single structural gene can give rise to different mature RNA transcripts, and then to different proteins. Once thought to be rare, alternative splicing is now known to be almost ubiquitous in eukaryotes, and in particular in humans (Wang et al. 2008). The production of alternatively spliced mRNAs is regulated by a system of trans-acting proteins, including splicing activators (which promote the usage of a particular splice site), and splicing repressors (which reduce the usage of a particular site). It is the cellular context, through intracellular and extracellular signals, which trigger proteins involved in alternative splicing. An important goal of current biological research is to determine exactly how proteins regulate alternative splicing, an idea sometimes presented as the quest for the “splicing code” (Barash et al. 2010; see also Matlin et al. 2005).
Though sequence activation and selection are very important phenomena, the most challenging process with regard to classical conceptions is sequence creation, by which the linear sequence of the final product is extensively scrambled, modified or created by a series of co- or post-transcriptional regulatory mechanisms. An example of sequence creation is trans-splicing phenomena, which are inconsistent with Crick’s sequence hypothesis “because they change in a regulated way the linear order of the elements in the product with respect to the order of the elements in the DNA from which those elements are derived” (GP, p. 93). A second example is RNA editing: RNA editing changes the primary sequence of mRNA during or after its transcription via the site-specific insertion, deletion, or substitution of one of the four nucleotides. A third example is translational recoding, where the message is recoded through frameshifting, programmed slippage or bypassing, or codon redefinition.
Griffiths and Stotz do an impressive job in presenting these phenomena in a clear and convincing way, despite their complexity. One can regret, nonetheless, that, probably because they wanted to remain as accessible as possible, they do not discuss some recent data that would contribute to strengthen their claims. For example, Griffiths and Stotz emphasize that a major aim of current molecular biology is to decipher the “splicing code”, that is, to uncover and predict how different mRNAs can be generated from a single piece of DNA, but, surprisingly, they do not mention the recent and striking progresses that have been made in this area. In particular, Barash et al. (2010) have used computer tools to develop a “splicing code” that enabled to predict how hundreds of RNA features work together to regulate tissue-dependent alternative splicing for thousands of exons. This “code” has been used to shed light on new biological mechanisms and to predict how alternative splicing is involved in different processes, including major developmental processes, and many of these predictions have proved correct (Fu and Ares 2014). An analysis of recent progresses made concerning the functioning of the proteome would have also been extremely interesting (see, for example, Cox and Mann 2011; Altelaar et al. 2013; Kim et al. 2014). It would also have been helpful to provide a discussion of prokaryote genetics, which is mentioned only “negatively”, meaning that it is contrasted with eukaryote genetics. This, of course, is related to the fact that gene regulation is not as complex in prokaryotes as in eukaryotes, but a discussion about how genes are regulated in prokaryotes and how gene regulation has evolved would have been useful. Naturally, a short book cannot cover all the aspects of today’s genetics, but it would have been sensible to bring up some recent results that make Griffiths and Stotz’s arguments even stronger.
Griffiths and Stotz gather the phenomena of gene activation, combination, and regulation under the label “molecular epigenesis” (following Burian 2004). “Epigenesis” has had historically several different meanings, but one of them is the antithesis of “preformationism” in the context of the embryology of the 18th century, and beyond: while preformationists considered that the final adult form is already contained in the egg (or sperm), epigeneticists asserted that the final adult form does not “pre-exist”, but is constructed progressively through interactions among all egg components and with the environment (Maienschein 2005). While many have criticized the recent resurgence of “preformationist” ideas in developmental biology (e.g., Oyama 1985; Griffiths and Knight 1998; Griffiths and Neumann-Held 1999; Lewontin 2000), some have suggested that a minimalist sense of “preformationism” was valid at the molecular level, in so far as, in virtue of the “genetic code”, the sequence of nucleotides on the DNA largely “mirrors” the sequence of amino acids on the protein (e.g., Godfrey-Smith 2000). In this context, the thesis of the “molecular epigenesis” defended by Griffiths and Stotz can be seen as a radicalization and extension of the anti-preformationist critique, with the explicit idea that preformationism is wrong even at the molecular level. Indeed, the different molecular mechanisms that regulate gene activity show that the correspondence between nucleotides and amino acids is, at the very least, much more indirect and “scrambled” than it was thought 30 years ago. Molecular “information” is not contained in any single molecule; on the contrary it is constructed through multiple interactions in processes regulated by the larger system at the level of the cell, the tissue or even the organism. Therefore, recent developmental and molecular biology suggest the validity of “epigenesis” at all levels, from molecules to individual traits.
The perspective of Griffiths and Stotz reflects very well the current move in genetics, which we can call the “epigenetization” of genetics. The focus on the mechanisms of gene regulation by RNAs and proteins has shown that there was no molecular genetics without molecular epigenetics, because to understand what genes are and what they do presupposes to untangle the complex mechanisms of that regulation. This is also the reason why genetics is in close and permanent interaction with the emerging fields of “transcriptomics” (the study of the complete set of RNA transcripts that are produced by the genome, under specific circumstances or in a specific cell) and “proteomics” (the study of the entire set of proteins expressed by a genome, cell, tissue or organism at a given time), the results of which are often published in journals specialized in genetics (e.g., Wang et al. 2009; Altelaar et al. 2013).
Thus, one can only hope that Genetics and Philosophy will contribute to convince philosophers of biology and, perhaps even more importantly, generalist philosophers that genes in the organism are not so much activators as they are activated, and that today’s genetics is interested in precisely how this activation is regulated.
Some of the weirdest genetic phenomena have very little to do with the genes themselves. True, as the units of DNA that define the proteins needed for life, genes have played biology’s center stage for decades. But whereas the genes always seem to get star billing, work over the past few years suggests that they are little more than puppets. An assortment of proteins and, sometimes, RNAs, pull the strings, telling the genes when and where to turn on or off. (Pennisi 2001: 1064).
Re-thinking developmental causality
On the basis of the above arguments concerning gene regulation, Griffiths and Stotz offer a rich reflection on biological causality and specificity, which constitutes certainly the most important philosophical contribution of the book. They defend that biological causality is both distributed and interactive. The emphasis here is on developmental causality. In fact, Griffiths and Stotz’s discussion about causality makes it clear that the main focus of the book is biological development, and more precisely how genetic and non-genetic mechanisms interact in making the development of an organism possible. A key question for several decades now has been to determine whether or not genes are causally special, in particular in development, and Griffiths and Stotz offer here their answer to this question.
To understand what GP brings to the debate over developmental causality, it is useful to step back, and say a few words about developmental systems theory (DST) (Oyama 1985; Oyama et al. 2001; Griffiths and Gray 1994, 2005; analysed in Barberousse et al. 2010). A key DST thesis is that many factors are involved in development, and these factors interact in complex ways. One crucial consequence is that a developmental resource becomes acause only in a given context, in which many other factors are involved. For instance, it is true that DNA is to some extent “already there” in an egg cell, but DNA becomes a cause only in an adequate cellular and extracellular context, for example because an hormone interacts with a protein, which in turn will activate one or several genes, which themselves will be involved in the synthesis of a protein. This is what Oyama (1985) has called, famously, the “ontogeny of information”, which means that even biological “information” itself does not just “lie” somewhere, ready to be expressed; even biological “information”, in other words, is constructed in the course of development, through complex interactions (see also Griffiths and Neumann-Held 1999; Stotz 2006; Noble 2006: 21).
In my view, this thesis about developmental causality, shared by several scientists (e.g., Lewontin 2000; Noble 2006, 2008) is important, and it is not trivial. Indeed, it is an arduous task to always keep in mind that developmental resources interact in complex ways, and that the cellular, organismic and environmental context significantly influences what genes can do in development. Some DST opponents have suggested that taking into account the intertwinement of developmental resources would just make research impossible in practice. How would it be conceivable, the objection goes, to pay attention to all the developmental resources and to all their interactions? No biologist could develop a research program based on such an encompassing view! DST would therefore be “holistic”, and “impracticable”, as famously suggested by Kitcher (2001). With the benefit of hindsight, Kitcher’s critique seems specious. First, the idea of “causal democracy” (all causes are of equal weight) is a straw man, not to be found in the writings of Lewontin or DST proponents (see Oyama 2000; Griffiths and Gray 2005; Griffiths 2006). Second, Lewontin and DST people’s view about causality reflected past empirical work ranging from genetics to developmental psychology (for example the work of Gottlieb and Lehrman: see Oyama et al. 2001; Griffiths 2006). Third, research done in the last 15 years (in particular about the epigenetic regulation of genes and about developmental symbioses, as detailed in the next section) has proved that biologists are increasingly aware of the importance of studying this intertwinement of developmental causes, and have developed remarkable tools to account for it. Thus, a general claim of DST, taken on by GP, is that there is no causal primacy of genes, and that genes become a cause only in a context.
But the initial question, “Are genes causally special?”, can come back under a different guise. In the last decade, a new attempt to defend that genes are causally special has been put forward by Waters (2007), and this view constitutes a new and crucial target of GP. Waters’ (2007, 572) aim is “to identify situations in which DNA is an ontologically distinctive cause, and to clarify the nature of its causal distinctiveness in these situations.” According to Waters, genes are causally special because DNA is the sole source of sequence specificity. Waters uses Woodward’s (2003, 2010) manipulationist account of causation, according to which there is a causal relationship between X and Y if it is possible to manipulate the value of Y by intervening to change the value of X. Woodward insists that the most important question is not to say if something is a cause, but rather to distinguish among causal relationships, according to some appropriate characteristics. One of these characteristics is specificity: a cause is specific if fine-grained changes in X lead to fine-grained changes in Y. For Waters (2007: 574), DNA is a specific difference maker because different changes in the sequence of nucleotides in DNA would change the linear sequence in RNA molecules in many different and very specific ways (in contrast, for example, with the way RNA polymerase influences the RNA linear sequence). In response to Waters, Griffiths and Stotz explain that they perfectly accept Woodward’s manipulationist account, and that from this perspective an important question is indeed to determine which causes are specific. But it is simply not true that DNA is the only specific cause. To demonstrate this, they propose to call “Crick information” (or “informational specificity”) the specific determination of the linear sequence of a gene product, in line with Crick’s (1958) proposal. They then show that several “epigenetic” regulatory mechanisms examined above (in particular trans-splicing, RNA editing and translational recoding) can exert a specific causal influence on the determination of the linear sequence of gene products. In other words, specificity is “distributed”, that is, shared among many different developmental factors. Thus, distinguishing specific causes is indeed important, but, contrary to Waters’ view, genes are not the only specific difference makers.
They show that the notion of Crick information (informational specificity) is useful in the context of molecular biology. Nevertheless, one should in fact adopt an extended informational specificity, according to which informational specificity is distributed among many developmental resources.
They express more sympathy with the idea that the “genetic code” can be understood as one key evolutionary innovation by which informational specificity is faithfully transmitted from one generation to the next (following Shea 2007, Bergstrom and Rosvall 2009).
Is the “developmental niche” equivalent to the “DS”?
Does the developmental niche expose itself to the same possible objection as the “DS”, namely that it is too inclusive and therefore imprecise?
Is development (rather than evolution) the main focus of a “developmental systems perspective”?
Is the organism an adequate level of analysis for this perspective?
One can regret that these questions have not been addressed in GP, and I look forward to seeing how Griffiths and Stotz situate their current perspective with regard to those questions.
In the last section, I turn to what I see as an important factual objection.
The missing final blow: developmental symbioses
Griffiths and Stotz examine, in Chapter 5, how environmental signals can influence the development of the organism, and, more precisely, how they impact gene regulation. They are also interested in how environmentally induced changes can influence the next generation. Examples discussed by Griffiths and Stotz are mainly related to the field of “epigenetics” understood in a broad way, and include genomic imprinting, cytoplasmic inheritance, and the role of some regulatory RNAs. Griffiths and Stotz also describe what they label “exogenetic inheritance”, including the inheritance of chromatin modification (Jablonka and Raz 2009). They also mention several research programs about environment-development interactions, including “ecological developmental biology” (Gilbert and Epel 2009), but only in passing (p. 130).
However, Griffiths and Stotz do not analyse in detail the phenomenon of developmental symbiosis, which is arguably the most innovating and convincing aspect of recent work done on how the environment influences development, and in particular on environmental regulation of genes (McFall-Ngai 2002; Gilbert and Epel 2009; Pennisi 2013; McFall-Ngai et al. 2013; strikingly, as early as 2002, in a special issue devoted to epigenetics, Gilbert (2002) was already insisting on the crucial importance of this phenomenon). This recent work illustrates what can be called the ongoing “symbiotization” of genetics, parallel to the “epigenetization” of genetics described above. For example, it has been shown in mice that a bacterial component of normal intestinal microflora modulates expression of host genes involved in key processes such as the maturation of the intestine and angiogenesis (formation of blood vessels) (Hooper et al. 2001; Stappenbeck et al. 2002), and that bacteria-induced postnatal organogenesis was frequent in the gut (Eberl 2005). Many other examples of how microbes decisively influence host oogenesis, organogenesis and morphogenesis have been documented across species (reviewed in Pradeu 2011). Several arguments make of developmental symbioses an extraordinarily convincing example to show that the environment can play a decisive role in development and how gene activation often depends on environmental factors. First, while some people might resist the idea that epigenetic regulation of gene activation in trans-splicing and similar examples is a form of environmental regulation, very few people would deny it in the case of microbe-dependent regulation. Incidentally, as Griffiths and Stotz note, some people are tempted to say that even epigenetic regulations can always be described as ultimately coming from the host genome, but this view is invalid in the case of microbe-regulated development. Second, the phenomenon of microbe-dependent development seems to be ubiquitous in nature, in plants, invertebrates and vertebrates (Gilbert and Epel 2009). Third, from an evolutionary point of view, some people deny that epigenetic regulations are important because they are rarely transmitted over many generations; I agree with Griffiths and Stotz that this is not a sound argument, as an influence over a single generation can perfectly be evolutionarily significant, but another striking feature of developmental symbioses is that they often involve microbes that are transmissible over many generations, with long-lasting effects, especially in arthropods (McFall-Ngai 2002), which makes the “multigeneration objection” invalid. Fourth, the effects of microbes on host development can be causally specific, in Woodward’s sense: in many cases, fine-grained changes in the composition of the microbiota can be related to fine-grained changes in host’s development or metabolism (e.g., Shin et al. 2011). For all these reasons, developmental symbiosis constitutes a major phenomenon to demonstrate the importance of environmental regulation of development, and in particular of the regulation of gene activation by environmental factors. It seems to me to be one of the most convincing examples of environmentally regulated “molecular epigenesis”, and one that obliges us to question the traditional boundaries of biological entities (Pradeu 2011, 2012). In my view, Griffiths and Stotz should have discussed in some detail this example, which seems so useful to strengthen their most fundamental claims.
GP is a very good and clearly written book. One minor difficulty lies in the many anticipations and reminders, with some pieces of information scattered in different places in the book. For example, detailed explanations about regulated recruitment and combinatorial control are scattered throughout Chapters 3 and 4, which forces the reader to go back and forth between these different sections. Another example concerns the description of epigenetic mechanisms, which leads the authors to mention in Chapter 3 Figure 5.1 of Chapter 5, a figure that indeed becomes fully understandable only in Chapter 5. Typos are very rare in the book, except perhaps for a minor confusion on page 68 between Science and Nature (Elizabeth Pennisi is a well-known editor of Science), a mistake (page 65) concerning the date of publication of Craver’s Explaining the Brain (2007, not 2009), and a misspelling of Denis Noble’s name (page 102).
The inclusion of some recent data on the local and environmental regulation of development (especially symbiosis) would have been helpful, all the more so as these data would have considerably strengthened Griffiths and Stotz’s point of view, in particular concerning molecular epigenesis and distributed specificity. Among other important advantages, these data confirm that an investigation of how different developmental causes are intertwined has become reality in today’s biology, against the “impracticality” objection of Kitcher (2001) and others.
Precisely because data in molecular biology change very quickly, up-to-date philosophy of molecular biology needs to change more quickly than other branches of philosophy of biology. It would be highly desirable, then, that Griffiths and Stotz plan updated editions of their book in the future and create a website with updated references and data.
It would have been extremely helpful, from a philosophical point of view, if Griffiths and Stotz had spelled out in more details what their main unit of analysis is: is it the organism? The “OE” (namely, the “organism-environment association”: Griffiths and Gray 1994)? The developmental system? If the later, what is the exact articulation between the “developmental system” and the “developmental niche”? This issue about the unit of analysis is related to an important debate, the respective role of the ultimate and proximate explanations in biology, and whether this distinction is legitimate. In my view, GP is about development and more generally about proximate rather than ultimate causes and it focuses on the organism as a major unit of analysis, but perhaps Griffiths and Stotz will disagree, and in any case I think that it would have been helpful to address these issues in this stimulating book.
- Beurton P, Falk R, Rheinberger H-J (eds) (2000) The concept of the gene in development and evolution. Cambridge University Press, CambridgeGoogle Scholar
- Brigandt I, Love AC (2008) Reductionism in biology. In: Zalta EN (ed) The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/reduction-biology/
- Burian RM (1995) Too many kinds of genes? Reprinted in Burian (2005), 166–182Google Scholar
- Burian RM (2005) The epistemology of development, evolution and genetics. Cambridge University Press, CambridgeGoogle Scholar
- Chen M, Manley JL (2009) Mechanisms of alternative splicing regulation: insights from molecular and genomics approaches. Nat Rev Mol Cell Biol 10(11):741–754Google Scholar
- Crick F (1958) On protein synthesis. Symp Soc Exp Biol 12:138–163Google Scholar
- Darden L, Tabery J (2009) Molecular biology. In: Zalta EN (ed) The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/molecular-biology/
- Falk R (2000) The gene: a concept in tension. In: Beurton PJ, Falk R, Rheinberger H-J (eds) The concept of the gene in development and evolution. Cambridge University Press, Cambridge, pp 317–348Google Scholar
- Fogle T (2000) The dissolution of protein coding genes in molecular biology. In: Beurton PJ, Falk R, Rheinberger H-J (eds) The concept of the gene in development and evolution. Cambridge University Press, Cambridge, pp 3–25Google Scholar
- Gilbert SF, Epel D (2009) Ecological developmental biology. Sinauer Associates, SunderlandGoogle Scholar
- Hull D (1974) Philosophy of biological science. Prentice-Hall Inc, Englewood CliffsGoogle Scholar
- Keller EF (2000) The century of the gene. Harvard University Press, CambridgeGoogle Scholar
- Kitcher P (2001) Battling the undead: how (and how not) to resist genetic determinism. In: Singh R, Krimbas K, Paul D, Beatty J (eds) Thinking about evolution: historical, philosophical and political perspectives. Cambridge University Press, Cambridge, pp 369–414Google Scholar
- Lewontin R (2000) The triple helix: gene, organism, and environment. Harvard University Press, CambridgeGoogle Scholar
- Maienschein J (2005) Epigenesis and preformationism. In: Zalta EN (ed) The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/epigenesis/
- Morange M (1998) A history of molecular biology. Harvard University Press, CambridgeGoogle Scholar
- Moss L (2003) What genes can’t do. MIT Press, CambridgeGoogle Scholar
- Noble D (2006) The music of life: biology beyond genes. Oxford University Press, OxfordGoogle Scholar
- Oyama S (1985) The ontogeny of information: developmental systems and evolution. Cambridge University Press, CambridgeGoogle Scholar
- Oyama S, Griffiths P, Gray R (eds) (2001) Cycles of contingency: developmental systems and evolution. MIT Press, CambridgeGoogle Scholar
- Sarkar S (2005) Molecular models of life: philosophical papers on molecular biology. MIT Press, CambridgeGoogle Scholar
- Schaffner K (1993) Discovery and explanation in biology and medicine. University of Chicago Press, ChicagoGoogle Scholar
- Stotz K (2006) Molecular epigenesis: distributed specificity as a break in the Central Dogma. Hist Philos Life Sci 28(4):527–544Google Scholar
- Waddington CH (1955) Mechanisms of development. Nature 4480(178):477–478Google Scholar
- Weber M (2005) Philosophy of experimental biology. Cambridge University Press, CambridgeGoogle Scholar
- Woodward J (2003) Making things happen: a theory of causal explanation. Oxford University Press, New YorkGoogle Scholar