Best way to quantify results from bacteria sample growth on petri dishes
November 22, 2016 8:16 AM Subscribe
3rd grade edition! My son grew bacteria on petri dishes for his science fair project. He collected samples from apples (unwashed, rinsed with water, rinsed with vinegar solution and water). The unwashed section grew fields of bacteria. The other sections are countable sometimes, sometimes with colonies. Does he just display photos or is there a way to graph or otherwise show the stark difference between unwashed and washed?
Photos are here if that's helpful (1-3 got contaminated and won't be used). Upper left: unwashed; upper right: water; lower right: vinegar; lower left: control. This doesn't have to be crazy intense because it's 3rd grade, but to me it looks like there isn't a major difference between water and vinegar, but a big difference between unwashed and washed with something. How does he best quantify or show this?
Photos are here if that's helpful (1-3 got contaminated and won't be used). Upper left: unwashed; upper right: water; lower right: vinegar; lower left: control. This doesn't have to be crazy intense because it's 3rd grade, but to me it looks like there isn't a major difference between water and vinegar, but a big difference between unwashed and washed with something. How does he best quantify or show this?
I was also going to suggest counting full and empty squares in each quadrant on a grid (just put the plate on some graph paper, or draw a grid over the photos you posted).
posted by pemberkins at 8:31 AM on November 22, 2016 [1 favorite]
posted by pemberkins at 8:31 AM on November 22, 2016 [1 favorite]
(Also, a good amount of that growth is likely yeast rather than bacteria, if you or your son happen to be interested in one versus the other.)
posted by pemberkins at 8:33 AM on November 22, 2016
posted by pemberkins at 8:33 AM on November 22, 2016
Response by poster: Point of clarification: we no longer have the petri dishes, only the photos. They got real stinky. We disposed of them in the proper manner.
posted by LKWorking at 8:36 AM on November 22, 2016
posted by LKWorking at 8:36 AM on November 22, 2016
Area's one way. Another is counting centre points.
If it was me, I'd probably count pixels (the logical end-point of counting squares) and do it in software by converting the image to a histogram.
You can create a histogram by hand, of course - just bucket the squares however you want, eg full/empty/partial or empty/white/yellow.
posted by Leon at 8:51 AM on November 22, 2016 [3 favorites]
If it was me, I'd probably count pixels (the logical end-point of counting squares) and do it in software by converting the image to a histogram.
You can create a histogram by hand, of course - just bucket the squares however you want, eg full/empty/partial or empty/white/yellow.
posted by Leon at 8:51 AM on November 22, 2016 [3 favorites]
Biologists usually count colonies, which means little mounds -- on a plate like #6, you essentially have a smear, so they can't be quantified, but most of the rest you could probably guess. If you want real quantification, you need to dilute your test sample to varying degrees to be sure that one of the samples ends up being countable, but here, I think probably "visual inspection" is plenty to say that unwashed is meaningfully different, and then you could do your best to quantify.
I suggest printing out the photos and using a sharpie to put a dot in the middle of each believable circle, and have a second person keep track. Given how few there are in the other groups, you probably can't prove that there are statistically meaningful differences among them, but the numbers might still be interesting (as might a discussion of what "significant" means).
Here's one stab, using plate #5, just to help you calibrate (I don't know how there are four letters for three conditions, but that's your problem, heh):
U: 60-70 (looking at screen, no sharpie)
W: 5
L: 0
V: 1
Honestly, these are pretty wacky because of the mix of critters being grown, but I think it would be fun to count them up...
Edited to explain: each "colony" represents one bacterium/critter that was present in the original sample, which then multiplies during incubation until there are enough to be visible as a mound. Small mounds, large mounds, and weird white mounds are each one family descended from a single starting cell, and the differences probably indicate different species/organisms. That's true to a first approximation, which is all you need. :)
posted by acm at 9:11 AM on November 22, 2016 [2 favorites]
I suggest printing out the photos and using a sharpie to put a dot in the middle of each believable circle, and have a second person keep track. Given how few there are in the other groups, you probably can't prove that there are statistically meaningful differences among them, but the numbers might still be interesting (as might a discussion of what "significant" means).
Here's one stab, using plate #5, just to help you calibrate (I don't know how there are four letters for three conditions, but that's your problem, heh):
U: 60-70 (looking at screen, no sharpie)
W: 5
L: 0
V: 1
Honestly, these are pretty wacky because of the mix of critters being grown, but I think it would be fun to count them up...
Edited to explain: each "colony" represents one bacterium/critter that was present in the original sample, which then multiplies during incubation until there are enough to be visible as a mound. Small mounds, large mounds, and weird white mounds are each one family descended from a single starting cell, and the differences probably indicate different species/organisms. That's true to a first approximation, which is all you need. :)
posted by acm at 9:11 AM on November 22, 2016 [2 favorites]
You need to go back down one step in quantification - how did you sample the bacteria?
You really have to relate back to # of cfu (colony forming units - the number of viable bacteria that will grow on your substrate) per unit of some kind, be it per total apple, per area of apple sampled, &c.
cfu is a practical way of assessing bacteria - even dead/non-viable bacteria can do harm, hence the caveat of cfu.
Typically, you would also do dilutions of the initial sample; the dilution scale will depend on your expected number of viable bacteria, typically this would be a 1:2 dilution, 1:10 dilution, 1:100 dilution.
When counting cfu, you'd normalize to the dilution factor and average the cfu out.
You also want to have positive and negative controls to demonstrate that your sampling method is valid. Typically, you'd grow up control bacteria to log phase then use OD400 to estimate the number of viable bacteria, then you do a dilution series as your positive control.
You'd also want to do a dilution series on sterile water to show that your agar plates and everything else you are using are uncontaminated.
The type of media that you use will also determine which classes of microorganisms you're sampling; typically there's total aerobic, total yeast and mold, and using even more selective media, you can reasonably identify different kinds of bacteria like salmonella vs.. e.coli vs. &c.
posted by porpoise at 11:17 AM on November 22, 2016
You really have to relate back to # of cfu (colony forming units - the number of viable bacteria that will grow on your substrate) per unit of some kind, be it per total apple, per area of apple sampled, &c.
cfu is a practical way of assessing bacteria - even dead/non-viable bacteria can do harm, hence the caveat of cfu.
Typically, you would also do dilutions of the initial sample; the dilution scale will depend on your expected number of viable bacteria, typically this would be a 1:2 dilution, 1:10 dilution, 1:100 dilution.
When counting cfu, you'd normalize to the dilution factor and average the cfu out.
You also want to have positive and negative controls to demonstrate that your sampling method is valid. Typically, you'd grow up control bacteria to log phase then use OD400 to estimate the number of viable bacteria, then you do a dilution series as your positive control.
You'd also want to do a dilution series on sterile water to show that your agar plates and everything else you are using are uncontaminated.
The type of media that you use will also determine which classes of microorganisms you're sampling; typically there's total aerobic, total yeast and mold, and using even more selective media, you can reasonably identify different kinds of bacteria like salmonella vs.. e.coli vs. &c.
posted by porpoise at 11:17 AM on November 22, 2016
Response by poster: The lower left quadrant in the pics is the control area; that's the 4th letter--C--that acm is referring to. I don't want to threadsit, but want to clarify I'm looking for the best way for my 8-year-old to show this information on a posterboard for his science fair project. The experiment is done and samples tossed. I think just showing pictures might be the best solution, though I appreciate the idea of doing a grid on the photos. Obviously, if my son's interest in this continues as he gets older his methods will become more scientific and involve more initial research!
posted by LKWorking at 12:39 PM on November 22, 2016
posted by LKWorking at 12:39 PM on November 22, 2016
Having judged science fairs, photos are fine. Most judges put more weight on the kid's explanation than the poster, ime.
If you're feeling fancy, you could download imagej (free!) and analyze the photos, it's easy to crop so you can select just one quadrant and there should be enough contrast to get decent results. There are lots of tutorials if you google, and it's something Real Scientists use.
posted by momus_window at 2:15 PM on November 22, 2016 [2 favorites]
If you're feeling fancy, you could download imagej (free!) and analyze the photos, it's easy to crop so you can select just one quadrant and there should be enough contrast to get decent results. There are lots of tutorials if you google, and it's something Real Scientists use.
posted by momus_window at 2:15 PM on November 22, 2016 [2 favorites]
Can confirm that Real Scientists use ImageJ--it's what I used to do all the quantification for my PhD thesis in biology :)
posted by Illuminated Clocks at 5:58 PM on November 22, 2016 [1 favorite]
posted by Illuminated Clocks at 5:58 PM on November 22, 2016 [1 favorite]
Best answer: Microbiologist here. I would come up with a simple scoring system. Something like 0 = no growth 1 = light growth (< 10 colonies) 2 = medium growth (10 - 50 colonies) 3 = heavy growth (50 - 100 colonies) 4 = too numerous to count.
posted by emd3737 at 10:42 PM on November 22, 2016 [2 favorites]
posted by emd3737 at 10:42 PM on November 22, 2016 [2 favorites]
Response by poster: Thanks for all the suggestions! The scoring system seemed to be the easiest solution and one my kid could accomplish on his own. Remember to wash your apples, everyone.
posted by LKWorking at 10:53 AM on November 28, 2016 [1 favorite]
posted by LKWorking at 10:53 AM on November 28, 2016 [1 favorite]
This thread is closed to new comments.
posted by btfreek at 8:28 AM on November 22, 2016 [7 favorites]