A 500 ml lab bottle with GL 80 opening was the main equipment for the reactor. Several holes were drilled into the cap and metal threads placed inside them.

Bacteria are moved into a UV-cuvette inside the measurement chamber of the fluorescence detector. The chamber is a 3D-printed enclosure (see Fig. Xy) with two walls facing each other that house a row of three LED sockets each. In the current setup, we only placed three [XXX UV?] LEDs in one wall. The other two opposing walls host camera sockets, to one one of which we have attached a RPi-camera, that is controlled by the RPi server. Two optical filters are placed inside. A [XXX] filter [do we show their spectral characteristics?] between the UV-LEDs and the cuvette, to filter out wavelengths from [XX] to [XX] nm (peak activation of eGFP at ~ XX nm). Between the cuvette and the camera a [XX] filter is inserted to allow only light emitted from activated bacteria through while holding back the rest of the LED-light.

The camera is controlled via the Python picamera library and calibrated to maximize the difference between idle and activated bacteria @allFromHof: (see table “camera-settings” or Supplemental Stuff, Zenodo, or nowhere?). The raw RGB-data is recorded, the green channel is kept and cropped to our region of interest---the bacterial suspension. The median green intensity of this region is then compared to a threshold [XX] to decipher the bacterial command.

3D-printed bla..

because of the expressed fluorescent protein’s spectral characteristics [explain shortly and link to molbio parts], we employ [XXX] nm LEDs to excite [XXX/proteins/chromophores?] and filter… images of LED’s emission spectrum and filter’s absorption curves for temperature experiment… ???

… see “colicam_nu.py” on github/daniel_moser/colibot/server/cam/ … 1 second shutterspeed..