Wet‐lab preparation and automated dataset decontamination procedure for bacterial single‐cell genomics

Abstract Applying single‐cell genomics (SCG) to microbes is an excellent opportunity to study their genome structure, diversity and evolution. The interpretation of SCG datasets depends directly on the quality of sample preparation and data processing, which should be as universal as possible to allow comparison and synthesis of the outcomes. We propose an adaptable and universal SCG library preparation protocol and an automated data decontamination tool, which we named SINgle‐CEll REads Decontamination through Automatic Taxonomic Assignation (SINCERE‐DATA) workflow. Our workflow identifies and removes contaminant reads based on (i) their contribution to high‐coverage regions of the assembly and (ii) their taxonomy at both read and contig levels. We applied these workflows to single cells isolated from laboratory strains and published SCG environmental datasets, demonstrating the procedure's efficiency in cleaning datasets. We highlight sources of contamination and sample preparation biases that jeopardise the quality of recovered single‐cell genomes. Using SINCERE‐DATA, we successfully removed contaminant DNA sequences from single‐amplified genomes that were undetected with the currently available decontamination tools used. In this work, we propose the full workflow embedding bacterial single‐cell library preparation and a dedicated bioinformatic tool, SINCERE‐DATA, to produce automatically decontaminated single‐amplified genomes. This workflow constitutes a new attempt to unify and improve the quality and comparison of the single‐cell datasets that will allow standardised examination of data contamination and provide cleaner genomes from the uncultivated majority.