Fig. 1: A brief overview of CellScanner is shown in box 1.

A classifier is trained on monocultures and blank data and subsequently applied to coculture data to classify every event ten times. When the repeated classifications agree on one species (or blank) for >70%, the event will be classified as such. Otherwise, the event will be classified as “unknown”. Box 2 summarizes the collection of mock community data. In short, monocultures were grown to stationary phase, cells were counted with a flow cytometer, and then diluted to obtain an equal number of cells per ml. These dilutions were used to make different ratios of the bacteria. Subsequently, mock communities were subjected to flow cytometry (FC) and 16S rRNA gene sequencing with Illumina MiSeq. Finally, FC data were analyzed with CellScanner.