Integration in biology and biomedicine

What it is, how it works, and what philosophers and scientists have to say about it

Integration is an increasingly important topic in biological and biomedical science. It refers to a range of activities, from combining methods and databases to merging explanations and disciplines. Although now emphasised in large-scale scientific endeavours, such as systems biology, integration has a history that can be traced back through a variety of fields and disciplines in order to understand the dynamics of integrative scientific practice. Integration has also been an important theme in philosophy of science. Although in the past philosophical discussion focused on the integration of theories, there are now several more broad-ranging philosophical accounts of integration. This workshop will bring together scientific, philosophical and historical perspectives on integration with the aim of developing new insights into the dynamics of integration in biology and biomedicine.


Please click here for a copy of the detailed program.

Program

Thursday May 3

10.00-12.00 Session I

Integration and its importance in contemporary environmental science
John Crawford (Sydney)

Soil science is a thriving field, driven into new modes of doing science by the proliferation of molecular data. Structural analysis, molecular quantification, modelling and experimental strategies are being brought to bear on soils in ever more integrated ways. Soils are increasingly being understood as self-organizing adaptive systems, with emergent properties that arise from interdependencies between physical, chemical and biological properties. As the interactions between microscopic and sub-surface biological agents in soil are better understood, dynamic insight into their effects on macroscopic above-ground organisms, particularly plants, has been gained. I will describe some of the theoretical work that has illuminated the dark world of soil systems, before going on further to look at the systems within which soil is embedded. In particular, I will argue that the great array of insights afforded to science by molecular and mathematical analyses cannot stop at the biological world: that in fact soil ecosystems must be analysed within the context of the social systems of which they are part. My general suggestion is that this must be the case for all system-oriented analysis of the biological world, and that strategies for reaching this level of system analysis must pervade the entire structure of research. I will outline a framework for how this could be achieved, and the challenges this poses for traditional structures of scientific research.

Obesity as a systems problem
Robert Dyball (Australian National University)

Obesity has reached epidemic proportions in many countries and persists despite continuing efforts to find solutions. Such "stubborn problems" often signal the influence of "feedback systems." In the case of the obesity epidemic, this possibility can be investigated using available system analysis tools. The investigation must begin with a study of the interplay between the full range of human and environmental factors. This paper outlines the nature of feedback and briefly discusses some of its management implications. A practical way to initiate a ‘systems approach’ to the obesity problem is suggested and four principles to guide the management of complex human– environment systems are presented.

Commentary: Stephen Simpson (Sydney)

12.00-1.00 Lunch

1.00-3.00 Session II

Integration from a philosophical perspective: past and present
Todd Grantham (Charleston)

This paper offers an overview of the primary ways philosophers have framed the topic of “integration” in the sciences. Although integration has historically been understood through the lens of reductionism, this approach is severely limiting: intertheoretic reduction remain “peripheral” to scientific practice; reduction focuses on the relations among theories to the exclusion of other aspects of science; and reduction fails to capture examples of non-reductionistic integration that have been scientifically important. For these reasons, recent work has tended to emphasise non-reductive conceptions of integration. Various non-reductionistic models of integration are reviewed (e.g., inter-field theories; explanatory extensions; integrated inter-level theories; hierarchies of mechanisms, etc.). These models address different “units” that can be integrated (theories, fields, explanations, methods, data, mechanisms). Whereas the philosophical question of reductionism is largely motivated by metaphysical concerns (physicalism, emergence, etc.), non-reductionistic models generally emphasise different aims: eliminating tensions among data/methods used by different fields and solving problems that cannot be addressed by the tools of a single discipline.

Explanatory integration and integrative pluralism
Sandra D. Mitchell (Pittsburgh)

Explanations of protein folding appeal to diverse factors including thermodynamic equilibrium, homology and the effects of intra-cellular molecular interactions. This case offers insights into how different perspectives which target different properties, use different methods and rely on different types of representation to display their results interact in practice to enrich our scientific understanding of this fundamental process. I will argue that integrative pluralism is the philosophical account that best captures the relationships among such diverse explanatory practices.

Commentary: Karola Stotz (Sydney)

3.00-3.30 Afternoon tea

3.30-5.30 Session III

Integrating molecular and systems approaches in chronobiology
William Bechtel (UC San Diego)

Provisional abstract: Across the life sciences researchers working on closely related phenomena (or even the same phenomenon) are often segregated into different fields or disciplines, publish in different journals and attend different conferences, and may be either unaware or incapable of relating to each others' work. Chronobiology, and especially the study of circadian rhythms, is a clear exception. Researchers from molecular biology, psychology, medical sciences and computational modelling are making productive connections to each others' work. I will argue that the frameworks of mechanistic explanation that requires both decomposing and recomposing mechanisms and dynamic mechanistic explanation that brings in computational modelling underlies this integration. As I will illustrate, accounts of circadian phenomena are multi-level, relating intracellular molecular components, intercellular and inter-organ coordination within an organism, and behaviours within social organisations on a planet that has a particular day-night cycle. Disrupting these at any level can result in serious affects on health.


Integration in systems biology and systems medicine
Olaf Wolkenhauer (Rostock)

Systems biology and systems medicine (which implements a systems biology approach in medical research) are not disciplines but areas of research that pursue an interdisciplinary approach by which biological and biomedical questions are addressed through integrating experiments in iterative cycles with mathematical modelling, computer simulation and theory. The aim of systems biology/medicine is to study how biological function emerges from the interactions between the components of living systems. More specifically, modern biomedicine seeks to explain the physiology, phenotypic behaviour or pathology of a tissue, organ or whole organism in terms of interactions and processes at the level of molecules and cells. I shall argue that the ultimate goal for these projects is the formulation of 'organising principles', robust generalisations that provide a deeper, more fundamental understanding of the behaviour of a system. This search for organising principles requires an integration of data, information and models – across technologies and across different levels of structural and functional organisation of a complex system. An analysis of the current practice in biological and biomedical research reveals that this cross-level reasoning involves de-contextualisation in order to generalise inferences. It can be argued that conventional approaches from Dynamical Systems Theory are not able to support this process and that for the search for organising principles new, more abstract, approaches are required to break out of the current pathway-centric framework and mechanistic modelling that dominates systems biology to this day.

Commentary: Sandra D. Mitchell and Ingo Brigandt

Friday May 4

9.00-11.30 Session IV

The complexity of complex diseases
David James (Garvan Institute)

Now that medical research has eradicated many life threatening infectious diseases and extended the life span of humans the next big challenge is to eradicate the more complex diseases of the modern world comprising cancer, diabetes, heart disease and neurological disease. Although these diseases have a genetic underpinning there is also an unquestionable link to the environment. Dissecting this interaction presents one of the major challenges of this century. I will present a view that to embrace this challenge we must embrace the complexity of biological systems. This requires a highly multidisciplinary approach involving collection of large data sets that capture the behaviour of the system as it undergoes key environmental transitions combined with analytic integrated approaches to reconstitute the system into a multi dimensional dynamic model that will rapidly facilitate prediction of system behaviour under a range of biological conditions.

Commentary: Steven Orzack (Fresh Pond Institute)

Systems biology and the integration of mechanistic explanation and mathematical explanation
Ingo Brigandt (Alberta)

I discuss how systems biology is working toward explanations that integrate mechanistic explanation and mathematical–two types of explanation that philosophers have viewed as excluding each other. Systems biology is an integrative approach in many respects, and it exhibits the combination of top-down and bottom-up approaches highlighted by recent philosophical accounts of explanation in terms of mechanisms. A clear virtue of philosophical accounts of mechanistic explanation is that they can cover multilevel and multified explanations. However, philosophers have advocated mechanistic explanation (as the breaking of a whole into structural parts and the description of their qualitative interactions) as being in opposition to traditional models of explanation as derivation from laws and equations. Philosophers' conception of a mechanistic explanation fails to capture several aspects of explanation found in systems biology. These include quantitative changes, feedback loops, emergent properties due to non0linear interactions, and system-wide properties such as robustness. Presenting examples from current systems biology I argue that mathematical modelling is an explanatorily relevant ingredient in the case of these phenomena, and that systems biology aims at complex explanations that integrate aspects of mechanistic explanation and mathematical explanation.

Commentary: Maureen O’Malley (Sydney)

Roundtable discussion: Integrating mathematical and mechanistic approaches

11.30-12.00 Morning tea

12.00-2.00 Session V

Data integration and community data-sharing
Sabina Leonelli (Exeter)

Each research tradition involved in data-intensive biology, and particularly the different cultures brought together through system biology, encompasses specific ways to select and evaluate data as evidence for claims about phenomena. The integration of such research traditions requires open discussion of the differences surrounding what counts as evidence and how data are used; and reaching some form of consensus, whether tacit or explicit, on how data are produced, selected, visualised and used to inform experimental and theoretical research. Consensus is often expressed in the choice and development of computational tools and digital databases to make data accessible and retrievable to a wide range of potential users. This paper examines the ways in which database development and use, and the related drive towards "machine learning," (1) embody the difficulties encountered when integrating data coming from different epistemic communities; (2) add to those difficulties by adding new levels of conceptualisation, representation and interpretation to the data being circulated; and (3) challenge the division of labour and accountabilities associated to data interpretation, by placing so-called "data scientists" (such as database curators) at the centre of decision-making processes about what should count as data in biology.

Open source drug discovery
Mathew Todd (Sydney)

Science shaped itself in the founding days of learned societies: individuals or teams competed, in secret, with paper-based communication in subscription journals. Why are we all still doing science like this? The internet has had a major impact in our sharing of data by traditional means, but it has not yet radically changed the way we actually perform science. Is competition between closed teams of researchers the best way to do science?
Our laboratory recently demonstrated the research acceleration possible when all data and ideas in a research project are freely shared. We solved a problem in the large-scale production of an important drug used to treat a tropical disease. In a new project we are attempting to find a new drug for malaria by operating the project openly on the web. There will be no patents. Anyone may participate - but will they?

Commentary: David Cook (Sydney)

2.00-3.00 Lunch

3.00-5.00 Session VI

Disciplinary integration and how to make it happen
Gabriele Bammer (Australian National University)

Despite considerable focus in "interdisciplinary" and "transdisciplinary" research on integrating knowledge from different disciplines and, quite often, from end-users and other stakeholders, there is still little practical guidance on how to do it. I’ve proposed three primary classes of methods: dialogue-based; model-, product- or vision-based; and common-metric-based. A small group of us published a toolkit of dialogue methods in 2009. We have had mixed success in seeking to expand it using an open platform. Another colleague and I are now working on a toolkit of modelling methods, which incorporates how different systems approaches bring knowledge together.

Using philosophy of science in interventions to facilitate disciplinary integration
Michael O’Rourke (Idaho) and Stephen Crowley (Boise State University)

Cross-disciplinary research is often collaborative, team research, due primarily to the need for a diversity of perspectives in addressing complex problems. Coordinating a team is famously hard work, but often overlooked amongst the logistical challenges are substantive conceptual challenges that are rooted in the very diversity of perspective that is key to solving these problems. Conceptual incommensurability can give rise to collaborative disintegration along disciplinary lines, which can manifest in communication breakdown and team collapse. One way to foster disciplinary integration, and thereby guard against damaging forms of miscommunication, is to develop mutual understanding about the conceptual assumptions that structure each team member's research contributions. While this understanding can arise organically out of the daily work of a collaborative, cross-disciplinary project, it can also be introduced through an intervention in the life of a team.
The Toolbox Project has developed a structured, workshop-based intervention that aims to produce this form of understanding through dialogue framed in terms of relevant concepts from the philosophy of science. In this talk we describe the development of our philosophically-framed approach to enhancing collaborative, cross-disciplinary research. We motivate the choice of philosophical framing and team-led dialogue, and then offer preliminary reasons to believe that this approach in fact facilitates disciplinary integration.
A key issue in developing this approach is identifying the relevant philosophical issues for the various kinds of teams. Standard philosophical techniques–both top down identification of relevant issues on the basis of theory and bottom up identification of issues through analysis of work in the disciplines–are employed in developing Toolbox instruments for different applications, such as translational health science. In addition to the obvious practical value of identifying relevant conceptual issues, we suggest that this way of working has the philosophical benefit of providing a novel way of testing philosophical hypotheses.

Commentary: Paul Griffiths (Sydney)

5.00-5.30 Closing remarks