BioLogicReactor
Introduction:
|
The BioLogicReactor is an incredibly cheap bio-reactor for doing experiments with bacterial, phage (and possibly yeast) cultures. By substituting computation and feedback for precision components and calibration, I'm reducing the cost of doing certain long-term (days to weeks) experiments in artificial evolution.
Cf. A System for the Continuous Directed Evolution of Biomolecules
I'm using a cheap USB camera to detect liquid levels in flasks, monitor turbidity in cell culturess, and (new) detecting bio-luminscence markers. To see the OpenCV-based level-detection at work, see: http://www.youtube.com/watch?v=46yMeK5wDpQ
Hardware:
- 1 Raspberry Pi
- 1 Logictech USB Webcam
- 1 Ardweeny
- 20 (or so) 3D printed parts (Clicking the link will download an MS Word Doc with part descriptions, images, and explanations.)
- 4 nylon septa
- 12 silicon tubing compression valves
- 1 valve actuator
- 2 motor mounts
- 4 Peristaltic pumps
- 3 Solenoid valves
- 1 Heat Gun
- 1 LM35-DZ Temperature sensor
- 1 10-Amp Solid-state relay
- 2 Styrofoam ice chests
- 1 Stainless-steel tray
- 1 Gearhead motor (Jameco 161382)
- 1 Gearhead motor (Jameco 155855)
Software:
Raspian (Wheezy) Linux
Arduino IDE
Python
OpenCV
SWI-Prolog
3d-printing biology engineering evolution microbiology
#Comments: 2
Thanks for posting this project on IGoR!
Really cool idea. I'd love to see it working.
I was thinking about using an approach like a Secchi disk to measure the turbidity of the bioreactors just using a camera, by placing a white and black pattern behind the fluid. Instead of measuring the depth at which the pattern disappears, you could measure an empirical relationship between turbidity and either the intensity difference the white and black portions of the pattern, or the sharpness of the edges. Edge sharpness might be more sensitive to turbidity, but would also be strongly affected by camera focus. One would probably need to also incorporate the intensity of the image of the white portion of the pattern to compensate for variation in light levels.
I may not be understanding the role of the bioluminescence correctly (from the attached file), but also I wondered if it would be easier to use some sort of color reaction instead of trying to measure bioluminescence. Perhaps culturing them with X-gal would give an assay for galactosidase production? Although using a reaction that pigments the fluid might interfere with turbidity measurements, one might be able to compensate by illuminating with white light and splitting the color channels (e.g. if using a blue pigment against a white background to measure protein, use the red channel to measure turbidity).