Thread by @yoginho

A little thread on the causal role of genes in complex regulatory systems, the subject of my recent paper in #BioEssays with philosopher James DiFrisco.

This is part of a collection of papers initiated by @drmichaellevin.

👇🏻

https://twitter.com/yoginho/status/1266418717988212736

@drmichaellevin Our paper introduces different views on causality to biologists. Physical notions based on law-like regularity of phenomena, or transfer of matter & energy are not really suitable in biology, where general laws are rare & the transfer of information often more important. /1

@drmichaellevin Instead, we use a notion of causality, (introduced by philosopher David Lewis) which is based on counterfactual conditionals, e.g. "if A would not have happened, B would not have happened, therefore A is a cause of B.” /2

@drmichaellevin A particularly useful counterfactual concept of causality for biology is developed in James Woodward’s interventionist account (“Making Things Happen”, OUP, 2003).

oxfordscholarship.com/view/10.1093/0… /3

@drmichaellevin On Woodward's account, A is a cause of B, if — when I intervene in a particular way on A — B will change. The methodology of classical genetics is obviously conforming very closely such an account: it treats genes as actual difference-makers, e.g. philsci-archive.pitt.edu/3833. /4

@drmichaellevin The special role of genes in biology has often been defended based on their causal specificity, i.e. their mapping to their effects in a fine-grained & well-defined way. /5

@drmichaellevin In our paper, we examine if such a privileged role for genes is still warranted in the context of complex regulatory dynamics. The answer is a clear no. /6

@drmichaellevin Feedback regulation & the multi-level nature of biological regulatory systems renders the causal effects of genes unstable, because they become heavily dependent on history & context. /7

@drmichaellevin Furthermore, individual genes often have no effect at all in redundant & robust regulatory systems, implying that they may not be proportional (at the right level of explanation) for such cases. /8

@drmichaellevin Finally, genetic effect are often no longer specific either in the context of complex regulation. Thus, genetic causation is unstable, no longer proportional, and unspecific at the systems-level. /9

@drmichaellevin To understand the causal structure of regulatory systems, we need to intervene on them at the systems level. This is only possible with properly formulated & validated dynamical systems models. /10

@drmichaellevin We show that inter-level causation through regulatory constraints also conforms to Woodward’s interventionist view of causation. To gain a causal view of a regulatory system, we do *not* need to reduce it to the molecular level. In fact, we need not to reduce it! /11

@drmichaellevin This account of genetic causation in complex regulatory systems shows that “bottoming out” at the genetic or molecular level is the wrong strategy if we are to understand causal effects on systems-level dynamics. /12

@drmichaellevin We propose that an alternative approach, based on Woodwardian interventions on constraints affecting the default dynamics of underlying processes, is more suitable to study the dynamics of regulatory systems. /13

@drmichaellevin I explain our analysis & our proposed approach in this lecture in my #BeyondNetworks course: . /14

@drmichaellevin Last but not least, I highly recommend James Woodward’s 2010 paper on “Causation in Biology” for definitions of causal stability, proportionality, and specificity: philpapers.org/rec/WOOCIB. /15