A user's guide to your first self-driving liquid handling lab

Experimentation is inherently difficult because most methods require substantial refinement, calibration, and validation before high-quality, reliable data can be collected. In most cases, experimental outcomes are impacted by multiple variables, thus requiring their simultaneous optimization for single and multi-objective targets. Traditional experimental approaches rely on trial-and-error methods guided by rational decision making, but these become increasingly inefficient and ineffective as complex interactions between inputs limit our ability to capture underlying trends using conventional statistical approaches. Machine learning and active learning (ML/AL) combined with automation represents an approach that can bolster future laboratory productivity. However, a steep initial learning curve and high costs of instrumentation pose substantial barriers to adoption. To democratize access, we herein comprehensively cover both the computational skills and hardware implementation necessary for self-driven experimental workflows. The accompanying open-source, low-cost liquid handling platforms offer practical templates for researchers adopting self-driving lab (SDL) methodologies. Complete tutorials and build guides are provided at https://gormleylab.github.io/SDLGuide.