[harrison]

Hello!

Yes, indeed, this was run in turbidostat mode, and using the “zigzag” function in the software.
You could set it up to do things differently. For example you could calculate approximate/rough growth rates from the pump rate data – i.e. doing a time average of this (and assuming OD is approxiamtely constant over time i.e. population is not net increasing/decreasing) you can get a rate of population growth.
Or you could even do a series of standard growth curves (without dilution) and fit exponential growth rates to these using any method of your choice (e.g. entirely separate to the Chi.BIo software, such as with R)

[harrison]

hmm. that may not work since the main loop of the experiment turns the pumps on and off each time it iterates (60 seconds). What you could instead do is put some logic check around the part in pumpmodulation which turns them on, and (for example) check if experiment cycle modulo 10 is 1 (or whatever) and if so turn it on, if not leave it off. This way you would let pumpmodulation run on the same timer but 9 out of 10 times it would do nothing.

[harrison]

Hello,
You could try that with the pump, but I am not sure it is the best idea. Why do you need to slow things down?
It sounds as if it would work though, you can easily test it in a dry run etc before actually tryhing in an experiment

[harrison]

– TO fix the flow rate thing, have a look at the user guide about this troubleshooting challenge – probably best method is to add a one way check valve to inflow media line.
– 5 times higher probably ideal. Or you could set it to do something like keep same rate as inflow then every ~10 minutes do a bigger pump out… But this would require some programming in python

[harrison]

Apologies for not being clear. What I suggest is start experiment. then DON’T click switch to activate OD control – since that is not a chemostat, but rather a turbidostat. Instead, once the experiment is started, go to the pump control section and enter a rate for each one and manually switch on the pumps. Then the system will run with constant inflow/pump rate (i.e. a chemostat).

Does this make sense?

[harrison]

Hello,

Using the device, you could just start an experiment then NOT turn on turbidostat mode, and instead permanently activate the in/out pumps at some fixed rate.
The turbidostat mode indeed does disable the punps – but for your sake you don’t need the OD control – hence you can do exactly as you did for a growth curve, and once it is running activate the pumps to whatever rate you want.
Make sure the outflow pump is much faster than the inflow so that it cannot overflow the reactor..!

[harrison]

Currently we use part 7757K41 from here:
https://www.mcmaster.com/products/one-way-flow-valves/check-valves~/body-material~plastic-1/check-valves-with-barbed-fittings-9/

[harrison]

Hello,
Are we talking about the culture tubes or the silicone tubing?
If the later, one thing to do is put a one way check valve between the pump and reactor (i.e. in the fresh media line). This prevents it moving backwards under gravity/siphon.

[harrison]

Hi Gabi, that is a tough one, sounds like a Beaglebone error/USB error rather something from our Chi.Bio operating system itself.
The fatal error often means the USB connection is not stable/was cut. Have you tried connecting it to a different PC and seeing if that solves it?

[harrison]

Hello
There is already an optogenetics mode which is basically doing this. It is not set up to run 10s every 5 minutes, but rather at variable intensities during each 1 minute cycle. Do you explicitly need the 5 minute cycle or are other ways OK?

In general if you want to implement from scratch, then yes, custom program is fine. I would look at the optogenetics one to use as a template for how to switch LEDs as and when needed.

[harrison]

Hello Sven
I don’t KNOW of anyone that has done this but It shouldn’t be too difficult. Main challenge will just be to write some python code that gets data into and out of the chi.bio back end. You could do this with http requests (i.e. the same mechanism currently used to update the GUI on the host computer). That should be fairly easy if you take a look at how the existing python (e.g. app.py) is structured.

[harrison]

ALas I dont believe it has been done. You would probably be best to just take a slow-motion video on a phone and then count rotations/time to get a rough estimate of RPM

[harrison]

The stirring system can be quite sensitive if it is a large stirrer with strong magnet, FOr this reason I would strongly suggest the recommended stirrer and running it at higher speed if you need better stirring.

Pump – It could be a hardware issue, have you tried a different pump board on the same reactor? If that works, but the othe rpump doesnt, likely that pump is broken.
For the OD function, the Pump=0.02 is the maximum rate. So, this is what you might expect it to deliver if you are way above the target OD. If you are delivering this amount of pumping but you are still unable to reduce the OD of growing cells, it could be the pump is not making a good seal on the tubing or for other reasons it isn’t ACTUALLY pumping. Look at the troubleshooting guide in the manual as there are various ways to improve this, e.g. putting some tape in the pump head, or putting a one way check valve between the pumps and the reactor so liquid cant flow backward at any time.

Harrison

[harrison]

Hello,
THe answer is there ARE differences between reactors and how important they are depends on the biological system + media + fluorescence signal + choice of excitation/emission bands. FOr example, each of the LEDs and each of the spectrometers have slight differences in manufacturing, and the “zero” level of the signal can depend very sensitively on these, so it is important to calibrate it out.

GENERALLY in our lab what we do is measure a “zero” signal for each reactor. For example putting the same non-fluorescent sample in each (i.e. just cells +media) , do the measurement, record the value as the “zero” value. Then, we subtract that zero from the experimental data to bring them to a common baseline. Doing this we find we can get very reproducible results, though admittedly it is much easier if you have a strongly expressed fluorescent protein and limited background (e.g. a non-fluorescent media)
Best wishes
Harrison

[harrison]

Yep, we have done this following that manual thing. If you do “cd //” can you see the etc directory? Or keep doing “cd ..” until you reach top directory then go down from there to find it 🙂

[Author]

Posts