Three weeks ago, on January 31st, students from the Urban Ecosystem stream in UT’s Freshman Research Initiative (FRI) met at their lab at 9AM and proceeded to collect water samples from Waller Creek at sites around Austin. These students belonged to two research teams within the stream: Naturalized E.Coli, and Molecular Source Tracking & Fecal Indicator Bacteria (MSTFIB). We sought to establish a baseline bacterial level for the creek, or in other words, figure out what the average bacterial levels were for the 12 sites we sampled. Some of our sites were north of campus (Hancock Golf Course), some were on campus (UT Medical School, Clark Field), and some were further downtown (Seventh Street, Tenth Street.) The diversity in locations allowed us to monitor any trends we could find up and down the creek and single out areas of interest with any really weird results.
After splitting into three groups, the samples were collected from 9 to 11 AM at all 12 sites. Two students with acid-rinsed sample bottles would go near the edge of the creek at each site, and would rinse their bottles three times in the creek water before taking a water sample.
After all the samples were collected, we re-grouped in lab and started plating our samples. “Plating” refers to moving small amounts of the creek sample to a petri-film/petri-dish, letting the plate sit overnight and observing all the bacteria that had grown overnight. The MSTFIB team specifically uses petri-films, which look like this.
The red dot with a bubble around it that you can see in the bottom right of the picture is a coliform (various bacteria) colony trying to grow in the nutrient-rich violet-red gel of the plate. The other red dots are irrelevant to our experiment. We like to pretend they don’t exist. They’re mostly used as a good benchmark of whether someone in our group knows what a bubble is or not (we’re still fighting over that definition.) There’s none in this picture, but a blue dot indicates the presence of E.Coli, which is EXACTLY what we’re looking for. E.Coli, as many of you know, is commonly found in the lower intestinal tracts of mammals. Not to go into too many details, but some unhygienic things needed to have happened for E.Coli to suddenly show up in the creek. According to the petri-film above, if you went swimming by Fifty-fifth street this last November 7th, you should’ve had a solid shot at avoiding high E.Coli levels. Can’t exactly say the same for the other thousands of bacterial species.
Back to the numbers. Our group wanted to reduce the effect of statistical error. One of the worst possible things in the world of stats is to have a sample size of 1. That means if we only made one petrifilm for each site, the results could get seriously wacky seriously fast. It’s possible, though unlikely, for a site with actually 1000 E.Coli cfu (colony-forming units) per 100 mL to produce a petrifilm indicating that there’s only 100 E.Colu cfu/100mL (which is very very false.) Therefore the importance of producing a variety of samples.
We collected samples at 12 sites. At those two sites, we took 2 samples each, one sample “A”, and one “B”. That’s 24 sites. Back in lab, from each sample, we created two petrifilms. One with only 1.0 mL creek water, and the other with 0.2 mL creek water and 0.8 mL PBS (phosphate-buffer saline solution.) That’s 48 petrifilms. Each petrifilm needs to be counted by two separate people to eliminate counting biases (which is great, because looking at these results, both of our wonderful counters can’t agree on how to count a handful of colonies and probably fight over the color of the sky too.) That means we have 96 total counts. Oh, and the E.Coli and coliforms need to be counted separately. That’s 196 bits of data that need to be entered into one table, and then averaged out. The poor soul who organized that table should get an award (thank you for your sympathy.)
Overall, this excursion was pretty fun and gave me tons of experience with sampling a variety of sites, multiplying its results to reduce error, and arranging data in easy to access ways. The MSTFIB team will use this data in upcoming weeks so we can analyze how bacterial levels will change in the days during and following heavy rain events. To do this, we will continue to take monthly creek snapshots, using this exact same procedure, hopefully streamlining the process each time while making sure our results are more precise. Our next creek snapshot is this Thursday, February 28th. Our data will allow us to see trends in bacteria levels across the creek, and if any sites increase or decrease a significant amount relative to other locations, we’ll make sure to pay a special visit to keep a close eye on those pesky fecal-indicator bacteria. You might be asking now, what’s the best part of working in this group? It’s obviously the color-coded data tables. There is no other correct answer.
Team Molecular Source Tracking & Fecal Indicator Bacteria
FRI Urban Ecosystems Systems