Thank you to David Jordan in Living Physics For support. All mistakes are entirely my own.
So, I tested 280 nm UV light on baker’s yeast.
It was a’‘Minimum Trust’ Investigating wet lab experiments, to learn the nuances of UV sensitivity data that I wouldn’t get from reading. My conclusion is that there are lots of sources of variability. This helped me examine for myself how delicate laboratory experiments can be in a real-world setting.
My main points were:
Calculating the UV dose applied to an organism is not easy. In my setup, I found it quite complex, with labs prone to error. I’m wondering if there should be a standardized UV dosage method within the community. When I read papers from others who calculate their doses, there’s a lot of math and a lot of assumptions. The amount of degrees of freedom here concerns me.
It is very difficult to sterilize your environment, and I really need to improve the wet lab setup I have at home. I also expect it to be a challenge for people doing independent science in general. So in my next post, I’m going to try to write more about sterile techniques for home laboratories, and think it will be useful for people doing any kind of biology experiment with less experience.
I think the benefit of doing independent experiments yourself is huge. I learned a lot about sources of variability in experiments that were not obvious to me at the beginning. One big thing that could happen is if this experiment shows no results. But, it happened and I’m going to study the first set of results. I’m lucky to be completely independent, so I don’t have to lie about niceties and I can say everything without worrying about it being successful.
Here are the yeast colonies that grew after different doses of UV light, shown in black text below each image. While the samples are contaminated, you can see that there is a Colony density decreases drastically as we increase the doseThe middle third sample got completely corrupted, which is why you can’t see anything there (I wonder why,,,),
At this point, I haven’t bothered to do exact calculations because I plan to repeat this experiment under better conditions, which I’ll write more about below. But, the initial phase looks promising. The first thing I find surprising is that the 1300mJ/cm^2 dose level did not completely eliminate the colonies, contrary to some papers I read that led me to believe that at this dose level I would be completely eliminated. But then, if you read on I will tell you why these data discrepancies might have occurred! ,
I made it first Correct concentration of yeast cells, using the previous serial dilution. Method in my previous post. Then, using a pump, I ran the solution through a quartz cuvette with 270-280 nm UV light (UV cannot penetrate regular glass). The setup is almost the same as I wrote about making it in my previous post automatic uv sterilization pumpExcept that I upgraded the pump so that it is connected to a motor driver. The motor driver lets me change the flow rate without messing with the duty cycles.
Here’s a refresher: From left to right, we have the relay that connects to the UV LED that is mounted in a frame that points toward the cuvette that contains the yeast solution. The black box with the blue circle is the peristaltic pump that pushes everything in. And the chip on the right is the motor driver, which is responsible for driving a variety of pulses that can activate the pump at different speeds.
The routine was as follows:
I dissolved my yeast and put it in the pump tube.
I set the pulse mode on my computer to program a specific pump rate.
I then measured how long it took the pump to dispense 1 ml of solution, this would be used to calculate the flow rate.
I then collected the solution that was pumped through the UV light, and then spread it on an agar plate with a cotton swab dipped in boiling water.
It was an emotional experience here. The experiment took about 3 hours because I was messing around with the test tubes, which made it quite difficult! It’s also annoying to sit through cleaning test tubes, repeating serial dilutions, making sure I don’t have a parallax error, etc. But overall, I enjoyed it. However, I can certainly see why novelty might be a bad thing for a researcher in the field.
Now, I am going to list here all the details that make these types of experiments a little difficult! This is an important part so please read further.
This is a little satire on the subtleties of UV dose calculation. Looking at my experimental setup, I found that it was quite difficult to actually calculate the UV dose I applied to the sample.
I found that estimating the actual dose delivered to the cuvette was very difficult, and the difficult part of the experiment. The UV passes through a quartz cuvette, and propagates through it, which I have modeled like a cone. The setup looks like this. The image on the left is of the front of the cuvette, with the rectangle window in the middle. On the right, I have shown what the UV ray from the bolt looks like. Hopefully I’ve described the kind of bizarre process required to estimate UV dose.
Ok, so how did I calculate the actual dosage given some flow rate? Here is a diagram of what the UV looks like, but rotated for clarity. I used technical data provided by the light manufacturer to estimate the intensity at the beginning and end of the cone.
To calculate the exact dose, I averaged the beginning and end intensities (but honestly this is a bit questionable, I should have done a weighted integral), calculated the cone volume, and then calculated the dose with the equations below. Note that this kind of thing depends on the assumption that the light manufacturer is actually declaring the correct intensity at those distances. This also assumes that I’m modeling the cone as some different shape – but that’s not the case because the intensity decreases as you go away.
Basically, it is difficult to measure the actual dose. For now, my preference is that these types of measurements may cause variability in experiments, as opposed to actual physical causes, but I need to think more about this.
That being said, here is an example of dose calculation for a given flow rate of 0.01cm^3/s.
In terms of gear, the actual UV light was an Ozonium UV Sterilization LED Bolt.
I used a peristaltic pump connected to the Pololu motor driver to control the flow rate. Motor drivers let you control the flow and Pololu has a GUI that makes it easy.
Peristaltic pumps work just like your intestines, squeezing and pinching, moving the solution along. Because of this, one must also take into account the subtle detail that water moves in a pulsating motion, not a continuous flow, which makes it even different from the scenarios in which you might actually think about using UV. This also adds some ambiguity to estimating the flow rate.
When running the pump, you have to calibrate the pulse rate to the flow rate, and the way I did this was by measuring how long it took 1 ml of solution to get into the pump. To do this I used gradations on my tube, which of course is subject to parallax/meniscus error.
Pump flow experiments are sensitive to contamination. In this experiment, contamination may come from the tubing, from the nozzle opening, from the cotton bud, or from any number of other sources. Therefore, for the next iteration I will use a more sterile technique, including an ethanol lamp to create a sterile field, as well as a flame burning the glass rod I am using to spread the agar plates.
This is something that is largely secret knowledge, and I really think more blogs about disinfecting, cleaning, and removing mundane mess would be really useful to independent scientists. I plan on filming some of these things with a body harness that can hold my phone as a camera, so stay tuned!