Spring 2022 Group 1

Introduction

Initially, our goal was to investigate how changing the expression level of certain genes affects memory capabilities in Drosophila presenting with Alzheimer’s Disease (AD), a condition which impairs memory abilities. However, most of the time was spent creating an effective testing assay (T-maze) to ensure accurate and applicable results. After discovering inconsistencies with T-maze results, we decided to alter the protocol by implementing aversive training.

Methods

Training Protocol:

In order to train starved flies to associate sucrose with a certain odor, we used a training protocol that involved exposure to both scents – one paired with agar and the other paired with agar and sucrose. The training involved letting the odor diffuse in all vials with the respective scent for 5 minutes, flipping the starved flies into the agar paired with the non-conditioned scent, transferring them into a resting vial, and then placing them into a vial with agar, sucrose, and the conditioned odor. After this was completed, they were placed in an agar only vial to starve again for 24 hours.

After varying the time for the odors to diffuse before training, T-maze odor concentration, training odor concentration and amount, amount of light, and presence of vacuum, we decided on the following setup. Please note that only one variable was changed at a time while finalizing this setup.

We prepped the training vials by using 0.8 µL of octan-3-ol (OCT) and 0.9 µL of 4-methyl cyclohexanol (MCH) diluted in 1.25 mL of white oil. After the solutions had mixed thoroughly and had been diluted, we placed 20 µL of MCH on filter paper, placed it on the filter cap of a training vial with agar only, and then placed tape over the cap. We did the same for OCT and placed it in a second training vial with agar and sucrose. We let the odor diffuse for 5 minutes. After the odor was diffused, we flipped 50 wildtype flies starved for 24 hours into the training vial with MCH and agar for 2 minutes, transferred them back into an empty rest vial for 2 minutes, and then flipped them into the training vial with OCT, agar, and sucrose for 2 minutes. We placed the training vials upside down when the flies were in the vials. After training was complete, the trained flies were placed into a vial with agar only to be tested the next day. The training described is shown in the first picture to the left.

Odor Balancing Protocol:

After discovering that there was not a clear preference for the odor paired with sucrose, specifically when OCT was paired with sucrose, it was decided to collect odor balancing data. We tested flies that were not trained to see if they do have a preference without training. This way we would be able to see if the T-maze resulted in a 50/50 split with untrained flies. Since they were not trained they should not have a preference for one odorant over the other. This protocol utilized the T-maze set-up/protocol with 50 untrained wildtype flies. It was then discovered that the flies seemed to inherently prefer MCH despite flipping the configuration of the odor vials on the t-maze.

Crosses: For the first half of the semester we set up crosses to collect progeny we needed to test. An example of a cross to setup is shown in picture 2. Before setting up crosses we discussed the experimental and control crosses we would need to setup.

Image of table we used to figure out what crosses we needed to setup from Nicole's notebook

T-Maze Design:

On the left: Original pictures from article. Materials, parts, and some dimensions are shown.

On the right: Our sketch rendering of the T-maze with dimensions.

T-Maze Protocol:

Our T-maze was set up with a diluted concentration of OCT on the right and a diluted concentration of MCH on the left. These diluted concentrations consisted of 1mL of the odorant and 2mL of water. Once the fruit flies were put in a plastic vial in the top o-ring of the T-maze, they were then transferred into the elevator. As the fruit flies were waiting in the elevator, an air-venting system was turned on to start the flow of both odorants. The fruit flies were then transferred from the elevator to the lowest o-ring. A one minute timer was set to allow the flies to make their preference towards either OCT or MCH. After one minute was up, we collected the flies in both odorant tubes to analyze which odorant accumulated the most fruit flies.

Results

What our graphs look like and how we represented our data.

This is the graph for the odor balancing tests that we did in the t-maze divided into days and orientations of the odorant vial OCT (right/left). The table provides specific preference indices as well as days these data points were acquired. We can see a strong preference for MCH in almost all of the trials.

This is a graph displaying the results of OCT paired with sucrose while keeping the concentration of MCH constant. We would have liked to see a clear preference for OCT at a specific concentration since it was paired with sucrose. There does not appear to be a clear preference for OCT from the data at any concentration level; indicating an inherent preference for MCH. Also, odorizing for only 5 minutes skewed the results as shown on 3/24/2022.

This is a graph displaying the results of MCH paired with sucrose while keeping the concentration of OCT constant. There is a clear preference for MCH no matter what concentration of MCH we used. As a group, we decided to use MCH as the control and to pair OCT with sucrose or arabinose for future experiments.

Conclusions

What conclusions can we draw based on our data and the relations we can create between different variables.

We came up with two main questions that we could answer with our data. These produced the main conclusions we could make utilizing the data we had collected throughout the semester.

Is the T-maze balanced?

This is the foundational question to ask because this is the foundation that we will be building all of our future data on. We need a balanced t-maze because then we have a standard to compare our training results to which tells us if we are actually training them to associate a certain scent with sucrose, compared to the flies having an inherent bias for one of the scents.
A sub-question that we also answered was: Does the configuration of the odorant vials (right or left) on the t-maze impact the flies' choice of scent?

Is there a significant difference in training effectiveness based on the concentrations of OCT and MCH?

Although both scents, OCT and MCH, are attractive to the drosophila, there is a certain point where that scent becomes aversive. We wanted to make sure our training was effective and not becoming aversive to the flies, so we had to control for this variable. We would also like to know how low our concentration can go to conserve resources and costs.

To answer these questions, we combined relevant data points from the graphs shown in the results section to create mean graphs and mean standard deviations and used an ANOVA to reveal if there was any significance to our findings. We defined significant as p ≤ 0.05.

The graph above shows the mean averages for OCT and MCH preferences for the simple odor balancing tests. We did not train the flies during these tests, so in theory, they should not have a scent preference. The table shows the specific data points used for the statistical graph.

This graph illustrates the flies' inherent preference for MCH and that orientation of the odorant vial on the t-maze has no significant effect on the preference of the odor. The MCH preference bar is higher in both orientations (OCT on right/left) meaning that flies preferred MCH despite us changing the t-maze configuration a bit. The standard deviations for the OCT preferences are quite low while the MCH preference bars' standard deviations are quite high. This means that OCT consistently had lower numbers of flies choosing it during the t-maze odor balancing trials while MCH had higher but fluctuating values for the flies that chose it during the t-maze trials.

Since the flies are not trained to associate a specific odor with sucrose, they should, in theory, have no preference and thus be evenly split for most of the trials. This was not the case according to our data, the flies revealed an inherent preference for MCH during the odor balancing trails. What this means for the rest of our results is that we have to factor in the fact that these flies seem to be more attracted to MCH. This means that when we are training them to associate sucrose with OCT, the results might not be as favorable as if we are associating MCH with sucrose due to that inherent bias. On the other hand, if we are able to train flies to associate OCT with sucrose, then that means that our training is effective and that we can train them to go against that bias. It means that our training is making a stronger association that can override inherent instincts.

The graph above is the mean averages of the tests using 7 minutes for the odorizing of the training vials for medium concentration of MCH and varying concentrations of OCT. It shows the error bars as well. The table shows the specific data points illustrated by the graph.

Based on our ANOVA and graph, we can conclude that there is no significant difference between the low OCT, med OCT, and high OCT because the p-values were not significant. This means that there would not be a significant difference if we utilized high, medium, or low OCT paired with sucrose in terms of preference for a scent. This is useful information because we now can pick and choose which one we want to use and if we want to minimalize costs, we can utilize the low OCT concentration for training. It also reveals that our concentrations (high, medium, and low) did not fall into the aversive concentration area. If we wanted to look at specifics though, we could look at the standard deviations of each of the concentrations and we would realize that the only concentration that has error bars within the positive preference index is the medium OCT. This could be used as evidence that medium OCT is the best concentration to utilize because the majority of the data points did not fall into that negative preference index. The medium concentration of OCT seems like the safest bet, despite there being no significant p-value differences between groups. The standard deviation is also significantly lower in the medium concentration of OCT suggesting that those results may be more robust.

One thing of note is that in these specific tests, we trained the flies to associate OCT with sucrose but some of our standard deviation bars fall into the negative area. This is most likely due to the flies apparent inherent preference for MCH as seen in the odor balancing graph. This means that some of raw data points for the t-maze test fell into a negative preference meaning that MCH was favored over OCT despite the flies being trained for OCT.

The graph above is the mean averages of the tests using 7 minutes for the odorizing of the training vials for medium concentration of OCT and varying concentrations of MCH. It shows the error bars as well. The table shows the specific data points illustrated by the graph.

Based on our ANOVA and graph, we can conclude that there is no significant difference between the low MCH, med MCH, and high MCH because the p-values were not significant. This means that there would not be a significant difference if we utilized high, medium, or low MCH paired with sucrose in terms of preference for a scent. Compared to the OCT Table, the MCH mean averages have a much smaller standard deviation meaning that they consistently preferred MCH in a specific range of preference indices. None of the standard deviations dip into the negative preference levels, so MCH is consistently preferred compared to OCT. This is most likely due to the inherent preference for MCH that we found in our odor balancing.

To answer our questions:

The main conclusions that we can draw based on the data we have accumulated are that the drosophila have a preference for MCH in our data and that the orientation of the odorant vials in the t-maze has no significant effect on the odor choice. We also know that the t-maze is not balanced as we have seen an inherent preference for MCH. We would need more data to see if these results are entirely conclusive, but based on the data we have acquired MCH is more preferred. Now we are not exactly sure if it is because the odor is more attractive to the flies or if the airflow within the T-maze is affecting the trajectory of their courses. That could be a possible future direction or variable that we need to control for. When the sucrose (positive stimuli) is associated with OCT, we can see there is a positive association, but it tends to be stronger when MCH is associated with sucrose. We also know that there is no significant difference in training effectiveness based on the concentrations of OCT and MCH. The p-values were much too large to be considered significant, so there is no difference in the concentration we use when training when it comes to preference indices.

What's Next?

Knowing that our t-maze is unbalanced is not entirely a negative thing. If we can confirm with more data that MCH is more preferred inherently compared to OCT, we can then build the rest of our data off of that conclusion. If we are able to see a strong preference index for OCT when we train the flies for OCT associated with sucrose, then we can be assured that our training is working.

Aversive training is one of the few things we considered during the last couple of weeks of the semester. We did six different trials using aversive training by pairing medium [MCH] with arabinose and pairing medium [OCT] with sucrose. The results we received were extremely positive, all tests coming out in favor of OCT. This differs from our previous testing because sometimes even when OCT was paired with sucrose, the flies were prefer MCH. This introduction of an aversive association allows for a stronger positive association with OCT due to the punishment-like effect arabinose has on the flies. It would be beneficial for us to use this on the next experiments that we do, so that we can optimize the positive and negative associations to get the results we want.

As we can see drosophila is very wild and they certainly do have a mind of their own. We definitely have to look into optimizing the T-maze. Changes in apparatus may be relevant. Another necessary change would be dividing each category of fly not just by sex, but by genetics. Many questions are to be asked starting with their ability to choose. One possibility to look into is to see if each sex and each type of genetically composed drosophila has a certain preference. The next thing to do in this important project is to test drosophila by more divided and specific categories.