Michael Dietrich, PhD

Professor and Chair, History and Philosophy of Science Department. University of Pittsburgh

Title: Inference and Automation: The Case of DNA Evolution

This presentation explores the impact of computer automation on the problem of modeling nucleotide sequence evolution. I will consider how what at first seemed to be a simple comparison between sequences became increasingly mired in empirical and methodological challenges of managing and making both models and methods.  I will outline how models of nucleotide substitution proliferated creating a problem of model choice that was then embedded in a computer program that automated that choice.  By analyzing how this model choice program was used, I will argue that computer automation allowed most of the program’s users to disregard the explicit methodological guidance offered by the program’s creators and other methodological experts.  This episode raises significant questions about the adoption of new methods, the duties of experts in science, and the impact of automation on biology.