A Software System for Automating Lab Experiments with Liquid-Handling Robots

Abstract

This thesis is concerned with the problem of automating liquid-handling robots in synthetic biology, which is especially relevant to facilitate experiments that are difficult or impossible to perform manually. Research tasks that are not highly standardized are still rarely automated in practice. Two main reasons for this are the substantial investments required to translate molecular biological protocols into robot programs, and the fact that the resulting programs are often too specific a lab to be easily re-used and shared. Recent developments of standardized protocols and dedicated programming languages for liquid-handling operations addressed some aspects of ease-of-use and portability of protocols. However, they either focus on simplicity but disallow complex protocols, or they entail detailed programming and require advanced skills and efforts from their users. The contributions of this thesis are the following. In Chapter 1, we introduce the problem and our aims and describe the hardware systems and biological methods used in the thesis. In Chapter 2, we describe Roboliq, a software system that (i) handles portable protocols employing a special data structure, high-level commands, and artificial intelligence methods, and (ii) facilitates a tight coupling of the design-execution-analysis cycle by supporting experimental design, simulation, and tidy data output. The chapter investigates the criteria for portability and shows how these can be fulfilled using the proper data structures and high-level commands. It also investigates a generic and flexible system for converting high-level commands to the low-level commands required by robots. In Chapter 3, we present three proof-of-principle applications for the reproducible, quantitative characterization of biophysical characteristics of a GFP variant; the experiments demonstrate the system's ability to handle experiments that involve complex pipetting, multi-day measurements and manipulations, and time-critical procedures. In Chapter 4, we present quality control experiments that characterize the pipetting performance of the robot; this is used for troubleshooting, error detection, and bias correction. The experiments are backed by a mathematical model, and we use MCMC simulation to enable the analysis of a couple complex experiments that are not amenable to traditional statistical methods. A selection of diagnostic protocols for evaluating various aspects of the robotic system and its equipment are also presented. Chapter 5 concludes our evaluation of Roboliq.