THE CONTEXT OF DISCOVERY OF DATA-DRIVEN BIOLOGY

My PhD dissertation aims (1) at reconstructing the structure of the context of discovery of ‘data-driven’ (big data, data intensive) biology and (2) at comparing it to traditional molecular approaches. Within the current debate in philosophy of science, ‘traditional approaches’ in molecular biology should be understood as the discovery and heuristics strategies identified by mechanistic philosophers such as Carl Craver and Lindley Darden. Therefore, key questions of my thesis are: what is the structure of discovery of data-driven biology? Is data-driven biology methodology different from traditional molecular approaches? The reason for doing such an analysis comes from a recent controversy among biologists. In particular, sides disagree on whether high throughput sequencing technologies are stimulating the development of a new scientific method somehow irreducible to traditional approaches. I will try to disentangle the debate by reconstructing and comparing data-driven and traditional methodologies. The dissertation is composed of five chapters. The first chapter deals with methodological issues. How do I compare data-driven and traditional molecular biology structures of discovery? Mechanistic philosophers have extensively characterized the discovery structure of traditional molecular biology. However, there is not such an analysis for data-driven biology. In order to do this, I will critically revise the discovery/justification distinction. The debate on discovery/justification has provided valuable tools on how discovery strategies might be conceived, and it is clearly one of the main forefathers of recent philosophical discussions on scientific methodologies in biology and physics. In Chapter 2 I shall to try to infer a full-fledged account of discovery for data-driven biology by means of the philosophical tools developed in Chapter 1. This analysis will be done in parallel to the investigation of key examples of data-driven biology, namely genome-wide association studies and cancer genomics. In Chapter 3 I analyze the epistemic strategies enabled by biological databases in data-driven biology. In Chapter 4, I will show how the discovery structure of ‘traditional molecular biology’ can be more efficiently rephrased through the same theoretical framework that I use to characterize data-driven biology. Since data-driven and traditional molecular biology seem to adopt the same discovery structure, one might consider the controversy motivating my research ill posed. However, in Chapter 5 I shall argue that there is still a valuable reason of disagreement between the sides. Actually, data-driven and traditional molecular biology endorse different cognitive values, which provide the criteria for evaluating models and findings as adequate or not. Here one might say that, although the structures of discovery (i.e. how reasoning and experimental strategies are structured and depend on each other) of the two sides are the same, the contexts of discovery (i.e. the set of both reasoning/experimental strategies and epistemic values/background assumptions that motivate discovery) are different. Therefore, in this last chapter I shall pinpoint the cognitive values behind traditional and data-driven biology, and how these commitments stimulate the heated disagreement motivating my research.