Rebooting Biology? Critical Reflections on Following Automations and Machine Learning in Synthetic Biology

• In Event: Lab Studies Reloaded? Machine Learning, Ethnography, and Critical STS

Abstract

Automations that researchers hope will make biological science faster, more reproducible and cheaper are finding their place among the existing epistemic apparatus of microbiology. Earlier laboratory studies were interested in what laboratories produce - inscriptions, enduring facts and technical apparatus. Later studies were interested in the disunity of science and demonstrated the differences between labs in different discplinary and national contexts. But laboratories were still understood as (re)producing cultures. To add to this body of research the project asked: how is it what goes into a laboratory, in this case machine learning, made useful and incorporated into scientific activities and agendas?

This paper reports on an ethnographic research project of an interdiscipline: synthetic biology. Here, computer scientists are deploying machine learning to combinatorially optimise biology by metabolic reverse engineering, changing experimental design and directed evolution. Amid these changes, new forms of doing research are emerging where timed, all-hands-on-deck ‘pressure tests’ are meant to demonstrate synthetic biology’s capabilities to engineer microbes that produce selected materials and chemicals. The paper discusses methodological issues arising from this research including: 1) the ‘point of entry’ and the analysts’ expertise in the entities being followed, for example: synthetic biologists (humans), machine learning (technology), and reproducibility (scientific norm)? 2) the effects of being ‘on site’ including modes and temporalities of intervention 3) what it means to study reproducibility in an ‘engineering’ rather than ‘scientific’ area of research 4) what researchers do to ‘classical’ lab studies as biology gets rebooted.

Author

• Robert Meckin, University of Manchester