The Duckbot: A system for automated imaging and manipulation of duckweed.

Abstract

Laboratory automation can boost precision and reproducibility of science workflows. However, current laboratory automation systems are difficult to modify for custom applications. Automating new experiment workflows therefore requires development of one-off research platforms, a process which requires significant time, resources, and experience. In this work, we investigate systems to lower the threshold to automation for plant biologists. Our approach establishes a direct connection with a generic motion platform to support experiment development and execution from a computational notebook environment. Specifically, we investigate the use of the open-source tool-changing motion platform Jubilee controlled using Jupyter notebooks. We present the Duckbot, a machine customized for automating laboratory research workflows with duckweed, a common multicellular plant. The Duckbot comprises (1) a set of end-effectors relevant for plant biology, (2) software modules which provide flexible control of these tools, and (3) computational notebooks which make use of these tools to automate duckweed experiments. We demonstrate the Duckbot's functionality by automating a particular laboratory research workflow, namely, duckweed growth assays. The Duckbot supports setting up sample plates with duckweed and growth media, gathering image data, and conducting relevant data analysis. We discuss the opportunities and limitations for developing custom laboratory automation with this platform and provide instructions on usage and customization.

Plant phenotyping relevance

植物研究向けの自動化プラットフォームを開発し、画像データ取得と解析を含むduckweed成長アッセイを自動化しており、表現型取得ワークフローが中心的です。

Code and data availability

The authors explicitly state that all design files, build instructions, example data (including the duckweed growth assay images) and analysis scripts are publicly available in their GitHub repositories, which directly reproduce this paper's duckweed phenotyping measurements and computational analysis.