Post #3: Zebrafish Lab Work Finished, Data Entry & Statistics have begun!

My three weeks of on-campus lab work have officially concluded! Now it’s time to go through all the data I collected, write a scientific-style paper about my results, and create a poster that I can display at the Undergraduate Research Symposium. I will talk more about my paper and poster in my summary post, for they are not yet complete, but I can describe what I observed in the lab during week 3 and the preliminary statistical results I have to share!

This past week, I conducted another round of experiments with 5 treatment groups–similar to my week 2 set-up–including a control, a “low” alcohol dosage, and a “high” alcohol dosage. Specifically, the “low” dosage groups were treated with 1.0% ethanol at 2 hpf and 4 hpf, while the “high” dosage groups were treated with 1.75% at the same time points. This allowed me to generate more data for the 1.0% treatment groups (as I had the same treatment groups in week 1) and gather information for a new dosage at 1.75%. These final two groups finalized my spectrum of ethanol concentration, which now includes 1.0%, 1.25%, 1.5%, 1.75%, and 2.0% treatments when combining all three weeks of data. This is excellent for statistical testing and analyzing bar graphs to determine (1) whether there are statistically significant differences between the control and treatment groups, (2) if there are any significant differences between treatment groups, and (3) if differences exist between groups, which groups are different from each other. The large amount of data also includes a wide range of my own observational notes, which were details I wrote in my lab notebook about the zebrafish in each treatment group at 24, 48, and 72 hours after they were born. These qualitative observations, in combination with the quantitative statistical and chart information, provide me with a solid amount of data to pull from when I start writing my paper and creating my poster. While the majority of my final data will be summarized in my last post, I was able to get through a significant portion of the data entry during my down time in week 3, so I can share some results with you all!

The data entry I finished was measuring blood vessel lengths of the zebrafish I photographed in weeks 1 and 2. I used an image analysis tool called ImageJ or FIJI, which includes a Neurite Tracer tool that is usually utilized to measure axons or dendrites extending from neurons. The tool works well for blood vessels too, so I was able to select the start and end point of each vessel, and FIJI would measure the length of the vessel in pixels. I compiled all of my data and ended up with 12 measured treatment groups to test! This includes all of the time points (2 hpf, 4 hpf, and 6 hpf) and ethanol percentages (0%, 1.0%, 1.25%, 1.5%, and 2%) from weeks 1 and 2. The first thing I did was create a bar graph showing the average ISV length for each group (ISV meaning intersegmental vessels; blood vessels that connect the main zebrafish artery, located on its back, to a vessel running longitudinally towards the belly/bottom of the zebrafish–this is just a more specific definition of the kind of blood vessels I was measuring). I then added error bars to get an initial idea of the kinds of differences that might appear in the statistical tests. My project advisor and I could predict that the experimental groups would be significantly different from the control based on the error bars, but it was difficult to tell any differences between treatment groups.

After running an ANOVA test, a p-value smaller than 0.05 was generated, which suggests that at least one of the tested groups is significantly different from the rest. I then ran a post-hoc Tukey test to determine what groups were different from each other. The results were that all of the treatment groups I tested–1.0%, 1.25%, 1.5%, and 2.0% at 2 hpf–were different from the control, as my project advisor and I predicted based on the error bars. 1.0% and 1.5% were significantly different from each other, and 1.0% and 2.0% were significantly different, but any other differences between treatment groups were insignificant. These results are extremely exciting, because above all else, the ISV length data metric we selected did demonstrate significant differences between all tested treatment groups and the control group! I will have to further analyze statistical information from other time points and between other treatment groups, but it is still very exciting that we were able to see differences even at low treatment doses.

I also began looking at bar graphs of the phenotype categories from week 2 and week 3 (normal, mild, moderate, and severe characterization counts for each treatment group), as well as heart rate, but I will need more time to run tests and look at error bars before I can make any conclusions about those data metrics. Those conclusions will be ready when I make my final summary post. Overall, I thought this final week in the lab was absolutely fascinating and eye-opening and the treatment groups I selected should round out my data well. I’ll be back for my last summary update after I go through the rest of my data, read through some scientific papers, and produce some final products of my own. For now, I’m gone fishin’ for statistical results and possible explanations for those results!


  1. Getting statistically significant results is always exciting! I remember thinking that our zebrafish experiment in intro bio was really cool, especially because it has health implications. Did you have a specific way that you determined phenotypic differences between fish? When I was in high school, I had to characterize worms by phenotype and that was definitely hard without a good definition that differentiated between phenotype classifications.

  2. casabochick says:

    Most definitely! And I completely agree, the health implications are part of what interested me and led me to select this project to begin with. In terms of phenotypic differences, I created four different levels of characterization–normal, mild, moderate, and severe. At each observational time point, I assigned each embryo to the level I felt best fit them. I had a list of characteristics that I used to help decide which level was best for each embryo, but it was very difficult! It was especially difficult at 24 hours past fertilization, because at that point many of the most distinguishing characteristics had not developed yet, and it was hard to decide between normal and mild or mild and moderate. Many quantification methods that rely on researcher classification, like my zebrafish phenotypes and your worm phenotypes, are difficult to use, but as more studies establish solid ways to differentiate, hopefully the process will become easier!

    Thank you for your comment!

    – Christina