Fig 2: A photograph of the experimental setup, inside the Opentrons OT-2 automatic pippetting robot. To the front left are the stock vials for aliquot additions and to the right is the reaction vessel wrapped in aluminium foil. The Opentrons OT-2 dimensions are 63 cm × 57 cm × 66 cm.

Fig 3: ReacSight-based assembly of an automated platform enabling reactive optogenetic control and single-cell resolved characterization of yeast continuous cultures.a Platform overview. The Opentrons OT-2 pipetting robot is used to connect optogenetic-ready multi-bioreactors to a benchtop cytometer (Guava EasyCyte 14HT, Luminex). The robot is used to dilute fresh culture samples in the cytometer input plate and to wash it between time points. The ‘clicking’ Python library pyautogui is used to create the cytometer instrument control API. Custom algorithms were developed and implemented in Python to automatically gate and deconvolve cytometry data on the fly. Two versions of the platform were assembled, using either a custom bioreactor setup (left photo) or Chi.Bio reactors4 (right photo). b Description of the gating and deconvolution algorithm. As an example, deconvolution between the overlapping fluorophores mCerulean and mNeonGreen are shown. c Stability of single-cell gene expression distributions over many generations. Strains constitutively expressing either mCerulean, mNeonGreen or mScarlet-I alone or altogether (‘3-colors’ strain) from the transcriptional units driven by the pTDH3 promoter and integrated in the chromosome were grown in turbidostat mode (OD setpoint = 0.5, upper plots) and cytometry was acquired hourly (vertical green lines). Distributions (smoothed via Gaussian kernel density estimation) of fluorophore levels (after gating, deconvolution, and normalization by the forward scatter, FSC) for all time points are plotted together with different color shades (bottom). RPU: relative promoter units (see Methods). The OD data for the ‘3-colors’ are not shown for simplicity and are similar to the others. d Characterization of a light-driven gene expression circuit based on the EL222 system17. Three different ON–OFF blue light temporal profiles were applied (bottom) and cytometry was acquired every 45 min. The median of gated, deconvolved, FSC-normalized data is shown (top). All bioreactor experiments presented in this figure were performed in parallel, the same day, with the custom bioreactor platform version. Source data are provided as a Source Data file.

Fig 4: ReacSight: a strategy to enhance bioreactor arrays for automated measurements and reactive experiment control.a On the hardware side, ReacSight leverages a pipetting robot (such as the low-cost, open-source Opentrons OT-2) to create a physical link between any multi-bioreactor setup (eVOLVER, Chi.Bio, custom…) and the input of any plate-based measurement device (plate reader, cytometer, high throughput microscope, pH-meter…). If necessary, the pipetting robot can be used to perform a treatment on bioreactor samples (dilution, fixation, extraction, purification…) before loading into the measurement device. If reactive experiment control is not needed, treated samples can also be stored on the robot deck for offline measurements (the OT-2 temperature module can help the conservation of temperature-sensitive samples). b On the software side, ReacSight enables full platform integration via a versatile instrument control architecture based on Python and the Python web application framework Flask. ReacSight software also provides a generic event system to enable reactive experiment control. Example code for a simple use case of reactive experiment control is shown. Experiment control can also inform remote users about the status of the experiment using Discord webhooks and generates an exhaustive log file.

Fig 5: ReacSight-based assembly of an automated platform enabling reactive control and characterization of bacterial cultures in low-volumes.a Platform overview. The Opentrons OT-2 pipetting robot is used to enhance the capacities of a plate-reader (Spark, Tecan). The robot is used to treat cultures in the plate-reader at predefined ODs. b Left: an E. coli clinical isolate can be maintained in growing conditions by renewing the media in an OD-controlled manner. Care must be taken to compensate for evaporation over extended time scales. Right: cells in rich media (glucose + casamino acids vs glucose alone) grow faster and yet resist better sub-MIC antibiotic treatments. c Left: A bacterial population may exhibit resilience to treatments thanks to the combination of two effects. At the single-cell level, cells may tolerate an antibiotic concentration exceeding their MIC through filamentation. Filamentation-based tolerance allows to increase biomass before cell death. At the population level, the antibiotic is degraded by enzymes released upon cell death in the environment. The final outcome depends on a race between cell death and antibiotic degradation. Middle: the respective role of these two effects can be investigated by means of repeated antibiotic treatments. Right: an E. coli clinical isolate is treated with different concentrations of CTX (legend) at an initial of OD of 5 10-4, and a second time with either 16 mg/L of CTX (red) or media alone (blue) at a user-defined OD (2.5 10-3 or 5 10-3). Because of instrument limitations, OD readouts below 10-3 are poorly reliable. Source data are provided as a Source Data file.