Fridolin Gross is a postdoctoral researcher who works on computational models of the cell cycle at IFOM, Milan, and teaches philosophy at the University of Kassel.
Occam’s Razor in Molecular and Systems Biology
Abstract
Biologists have been debating for some time about what — if anything — is conceptually novel about systems biology when compared to the ‘classical’ approach of molecular biology. In these debates scientists make use of concepts, such as holism, reductionism, or emergence, that have a long tradition in philosophy. Thus, philosophers might productively contribute to the discussion by taking into account this tradition and by adopting an impartial perspective with respect to the two competing positions.
In my talk I focus on one particular way in which the difference between molecular and systems biology has been spelled out. Some systems biologists have argued that molecular biology is misguided because it relies on an unjustified application of the principle of Occam’s razor. Roughly speaking, Occam’s razor refers to the idea that among competing explanations the simplest should be preferred. However, a look at the history of philosophy shows that this principle has been understood and defended in very different ways. I analyze which version of the principle is relevant in the present context and ask whether the allegation stands up to scrutiny when looking at instances of actual research in molecular biology. I defend molecular biology by arguing that its use of Occam’s razor is largely innocuous and by showing that systems biology heavily relies on considerations of simplicity as well.
Charles Pence is a Philosopher of biology at the Université catholique de Louvain, Belgium.
His work spans two distinct areas. First is the philosophy and history of evolutionary biology – in particular, the conceptual structure of evolution by natural selection and the reception of Darwin in the late 19th and early 20th centuries. Second, he is interested in the application and ethics of contemporary technology, both in the digital humanities and as applied to contemporary warfare.
“The Wonderful Form of Cosmic Order”: Bringing Statistics to Evolution
The introduction of statistical reasoning into evolutionary theory – and the corresponding changes in our concepts of chance and evolution that made it possible – did not happen overnight. In particular, many of the relevant biologists between 1900 and 1930 (such as W. F. R. Weldon, Karl Pearson, and R. A. Fisher) pointed back to Francis Galton as a vitally important precursor for their work. When we read Galton, however, it is far from immediately obvious why he would be taken to be so crucial by so many important successors. In this talk, I’ll explore some of Galton’s ideas about the relationship between micro-level variation and population change, and try to reconstruct what within his thought might have been seen to be so groundbreaking by biologists in the first part of the twentieth century. I will close by drawing some morals for our contemporary conception of evolutionary theory.
During his stay in Bordeaux Charles Pence will be available for discussions. If you would like to talk to him, please sign up here so that we can arrange for a meeting.
What is a stem cell? This is both a biological and a philosophical issue. As a philosopher, my aim is to describe the kind of property (or set of properties) that stemness is. On the basis of scientific literature analysis, interviews of stem cell and cancer stem cell specialists, and my own experience of wet lab experiments, I will address four successive questions:
(1) What kind of property is stemness? I will show that stemness can be four kinds of properties: categorical (stemness is an intrinsic property of stem cells), dispositional (stemness is an intrinsic property whose expression relies on extrinsic stimuli), relational (stemness is an extrinsic property induced by the microenvironment), or systemic (stemness is an extrinsic property that is maintained and controlled at the cell population level). Stemness is different in different types of stem cells (e.g. stemness is a dispositional property for the hematopoietic stem cells and a relational property for the germ line stem cells).
(2) Does it matter? I will show that clarifying the nature of stemness for each kind of stem cell is of major importance at least for oncology, as the efficiency of some therapeutic strategies (cancer stem cell-targeting therapies, niche-targeting therapies) directly relies on whether stemness is a categorical, a dispositional, a relational, or a systemic property (Laplane HUP 2016, reviewed in Clevers Nature 2016). In addition, the different types of stemness have specific consequences for evolutionary processes in somatic cells (Laplane Biology & Philosophy 2018).
(3) Is stemness stable? I will explore malignant hemopathies and show that some data suggest while stemness is a dispositional property in the normal hematopoietic system, it might switch to a categorical, relational, or systemic property, depending on genetic and epigenetic alterations (Laplane & Solary M/S 2017).
(4) Is stem cell a unified biological category? If stem cells fall under four distinct categories then are we giving the same name to unrelated cell types? I will suggest that crossing phylogeny and philosophy of stem cells might shed light on the natural boundaries of the concept of stem cell.
Refrences
Laplane L. Cancer Stem Cells: Philosophy and Therapies. Harvard University Press, Cambridge (MA), 2016
Clevers H. Cancer therapy: defining stemness. Nature 2016; 534(7606): 176-177
Laplane L et Solary E. Identité des cellules souches normales et cancéreuses. M/S 2017 ; 33 : 899-904
Laplane L. Cancer stem cells modulate patterns and processes of evolution in cancers. Biology and Philosophy 33: 18.
Lucie Laplane is a CNRS researcher at UMR8590 (IHPST), Université Paris I-Panthéon-Sorbonne & UMR1170, Institut de Cancérologie Gustave Roussy. Slides of Lucie Laplane’s talk.
Angela Potochnik is an Associate Professor of Philosophy and the Director of the Center for Public Engagement with Science at the University of Cincinnati. Her research addresses the nature of science and its successes, the relationships between science and the public, and methods in population biology. She earned her PhD from Stanford University in 2007.
Abstract
Across biology, attention to causes—or mechanism—is common, as is debate about which causes are important. In this talk, I motivate the idea that scientists focus not just on identifying causes, but on the identification of what I call causal patterns. This alternative framing is an important shift, and it alters what is at stake in the debate about when a causal account succeeds and in what this success consists. Focus on a given causal pattern is motivated by a combination of causal facts and the nature of our research priorities. This emphasis on causal patterns thus also changes how we should think about theory development. Theory change can be motivated by the discovery that a causal pattern fails to obtain, or by a shift in research priorities toward a focus on different causal pattern(s). This view predicts the maintenance of multiple accounts of the same phenomena, accounts that seem incompatible but that capture different patterns.
Actionable Data for Precision Oncology: Building Trustworthy Evidence for Exploratory Research and Clinical Diagnostics
Abstract
Difficulties in managing the enormous amount of relevant data being produced by researchers around the world continue to undermine data-centred discovery and therapeutic development. This is particularly evident when looking for evidence to identify which entities should be targeted, and how – an issue typically referred to as ‘actionability’ of data by practitioners (Nelson et al., 2013). This paper, co-authored with Niccoló Tempini from the University of Exeter, considers how researchers make decisions about the actionability of specific datasets and the extent to which such data can be considered to be trustworthy. To this aim, we discuss the case of COSMIC, a leading data infrastructure in cancer genomics which aggregates a large amount of data sources, ranging from literature to screen repositories and functional information about cancer at various levels of description (including gene, individual mutation, methylation and drug resistance). COSMIC occupies a central position in the path towards precision oncology. It is used by many research groups both for exploratory analyses and in drug or diagnostic development pipelines, and it has a strong reputation as a reliable source of genomic evidence for clinical use. On the basis of qualitative research on COSMIC data curation and use practices carried out between 2015 and 2017, we identify significant differences in the understanding of data actionability underpinning exploratory research on biological mechanisms as opposed to research aiming to develop diagnostic knowledge and related instruments/markers. Though sometimes inhabited by the same individuals, the exploratory and the diagnostic research spaces pose different requirements and constraints upon data use, which in turn shape different imaginaries of actionability and perceptions of what makes a data source trustworthy. We show how different ways to construe data use in medical research are integral to considerations of evidential reliability, and offer a definition of actionability and trust in data that is dependent on the medical goals at hand in any one situation of inquiry.
Do we need an extended evolutionary synthesis?
Significant advances in several areas of evolutionary biology have provided a wealth of new knowledge about the factors responsible for evolutionary change, prompting a debate about whether a renewed theoretical synthesis is required. This lecture will present the principal arguments in support of one of the proposals for a reformed theoretical framework, the extended evolutionary synthesis. Whereas the standard theory and its various amendments concentrate on genetic variation in populations, the extended framework emphasizes the role of constructive processes, causal reciprocity, and systems dynamics in the evolution of organismal complexity. I will discuss the principal components, logical structure, and specific predictions of the extended synthesis concept.
Alvaro Moreno Bergareche is Full Professor of Philosophy of Science at the University of the Basque Country (UPV/EHU) (Spain). Founder of the Philosophy of Biology Group at the UPV/EHU, he is a specialist of Philosophy of Biology, Artificial Life, Complex Systems and Cognitive Science.
Abstract
In this talk I will discuss individuality from the perspective of its origins. I will argue that the different forms and levels of biological organization ultimately date back to –and lie in– the prebiotic appearance of a minimal form of individuality, understood as a self-encapsulated self-maintaining (autopoietic) organization, which constitutes the core of what we mean now by organismality. Thanks to the spatially and temporally wider scaffold they created, these instances of actual organization have enacted a new world where they deal with a population of semi-autonomous genetic entities. This has led to the appearance of a new form of individuality that can survive and thrive incorporating immune mechanisms. Thus, at the end of the prebiotic evolution we have hierarchically and cohesively organized individuated systems which, closely associated with similar individuated systems, generate new cohesive associated entities. Finally, I will discuss how the earlier forms of life have been developed different forms of individuality and the limits of this process.
Several authors have argued that causes differ in the degree to which they are ‘specific’ to their effects. Woodward has used this idea to enrich his influential interventionist the- ory of causal explanation. Here we propose a way to measure causal specificity using tools from information theory. We show that the specificity of a causal variable is not well de- fined without a probability distribution over the states of that variable. We demonstrate the tractability and interest of our proposed measure by measuring the specificity of coding DNA and other factors in a simple model of the production of mRNA.
Based on: Paul E. Griffiths, Arnaud Pocheville, Brett Calcott, Karola Stotz, Hyunju Kim, and Rob Knight, Measuring Causal Specificity, Philosophy of Science (2015). Full text available Details on Arnaud Pocheville’s research
A philosopher of science with a focus on biology and psychology, Paul Griffiths (University of Sydney, Australia) was educated at Cambridge and the Australian National University, receiving his PhD in 1989. He heads the Theory and Method in Bioscience project node of the Charles Perkins Centre, a major new initiative at Sydney focused on interdisciplinary research into obesity, diabetes and cardiovascular disease. He served as Associate Academic Director for Humanities and Social Sciences while the Centre was being established, and continues to serve on the Executive Committee as Domain Leader for Society and Environment.
Abstract
Neo-Aristotelian theories of the organism have had a significant revival in recent years, with applications in ethics and in the philosophy of medicine. They have been criticised on methodological grounds, and for being unable to deliver plausible normative results. Here it is suggested that their characterisation of the telos (intrinsic purpose) of an organism is wrong in the straightforward way that someone who supposes washing machines are for cooking food is wrong. Organisms are not designed to reproduce their form, either ontogenetically, within an individual lifetime, or phylogenetically, across generations. An analysis of the design of organisms guided by this teleological heuristic will inevitably fail. Other, more naturalistic theories of the organism, such as those presupposed in ‘organisational’ accounts of biological function, make the same mistake. Only a fundamentally unintuitive account of organismic design, informed by our best current evolutionary theory, can make sense of many aspects of the organisation of living systems.
Marie Darrason est à pneumologue et philosophe de la médecine. Elle est interne de médecine des hôpitaux de Paris et Ancienne doctorante de l’Institut d’Histoire de Philosophie des Sciences et des Techniques (Paris 1/ENS/CNRS).
Voir son profil et ses publications.
Résumé :
L’essor récent des thérapies ciblées, fondé sur le concept de mutation actionnable, a fait émerger à nouveau le rêve de classifications moléculaires des maladies.
En effet, une thérapie ciblée est dirigée contre une altération moléculaire (mutation, réarrangement, amplification, etc) dite “actionnable”, c’est à dire une altération qui est essentielle au développement de la maladie et dont l’activation ou l’inactivation vont avoir un effet spectaculaire sur le cours de la maladie. Ainsi dans l’exemple du cancer du poumon, on distingue à présent plusieurs types de mutations actionnables (les adénocarcinomes EGFR mutés, ALK réarrangés), auxquels correspondent plusieurs types de thérapies ciblées (les anti EGFR, les anti ALK), qui permettent des réponses tumorales remarquables.
Dès lors, il est tentant de conclure que l’avènement des thérapies ciblées correspondrait à une redéfinition moléculaire de la maladie : à chaque phénotype particulier correspondrait une mutation actionnable et une thérapie ciblée spécifiques. Loin des distinctions anatomocliniques et sémiologiques grossières, il serait enfin possible de reclassifier une même entité chimérique en plusieurs maladies distinctes, de délimiter au sein de ce qui était considéré comme une seule maladie une multitude de phénotypes auxquels correspondraient des altérations moléculaires spécifiques.
L’objectif de ma présentation sera de discuter la réalité et la pertinence de cette redéfinition moléculaire de la maladie au prisme de l’analyse du concept de « mutation actionnable ». Je soutiens ainsi que le concept de mutation actionnable est un concept pragmatique, fondamentalement dynamique et qui est amené à se modifier dans le temps et dans l’espace, comme le montrent les phénomènes d’hétérogénéité intra et inter tumorale et l’émergence quasi inéluctable des résistances aux thérapies ciblées. Plutôt que de proposer une classification moléculaire de la maladie, je propose donc d’explorer le concept de trajectoire évolutionnaire de la maladie, qui me paraît particulièrement pertinent en oncologie médicale. J’essaierai notamment de distinguer le concept de trajectoire évolutionnaire de celui de signature moléculaire et je discuterai l’intérêt d’un tel concept hors du domaine de la cancérologie.
