MARCH 2022
Kyra's Blog
Kyra Bankhead, undergraduate student
1 March 2022
I had the privilege of meeting with Grace Freeman to talk about data analysis thus far. She had a lot of great points. First she had the idea of using only the first noise measurement of each survey because this is when the counts of the harbor seals hauled-out took place. She also suggested that tide level and temperature might be collinear with month and that perhaps both of these predictors should be taken out of the model. I was very confused why using the negative binomial model for the Waterfront resulted in a significant correlation with noise and number of harbor seals hauled-out. However, Grace explained to me that in the model with temperature, month, and tide level, with these constant, noise has an effect on the number of seals hauled-out. All of these are great places to start in improving my models for both sites.
Grace is going to start from scratch with the data to see if she gets to the same place as me. I am excited to see what she gets with starting over because I have been working on data analysis for two years now.
Until then,
Kyra Bankhead
Madison's Blog
Madison Gard, undergraduate student
1 March 2022
February flew by! Outside of the lab, I survived two rounds of midterms and managed four Discovery Day events for the Office of Admissions. Every day was filled with to-do lists and long hours of studying or drafting emails. Needless to say, I’m happy to see the quarter starting to wind-down. I’m so looking forward to visiting my happy place in Cholla Bay, Mexico over Spring Break. :-)
Lab itself was also a whirlwind. We started off strong and ran several plates of samples, though our results using the harbor seal Taqman ZFX and ZFY probes on the Steller and California sea lion samples weren’t confidence inspiring. We had trouble with the majority of sample wells on several plates not being detected by the qPCR instrument. To troubleshoot, Zoë and I tried several things. Our first guess (and hope) was that our master mix used for the amplification of target DNA regions was old and perhaps less effective, so we ordered a fresh batch. Sadly, the fresh mix didn’t make much of a difference in improving signal detections (it can never be the easy fix, can it?).
We had two other suspicions for why the sample detection rates were so low. The first of these is the quality of DNA in the aquarium scat samples could be too low to amplify, and the second is the harbor seal ZFX and ZFY probes are not a close enough match to the corresponding regions in sea lion genomes. This led us down a rabbit hole of a day attempting to find stretches of base pairs within the ZFX, ZFY, and SFY regions that matched in all three species: harbor seals, Steller sea lions, and California sea lions. We did this to design a new Taqman probe that would hopefully be a better match and more effective at amplifying the fluorescence signal for the qPCR instrument to detect. I really have to credit Zoë for walking me through this process conceptually and ultimately being the one to design our new Taqman probe, which we ordered last week.
The probe arrived last Thursday, and after two test runs there doesn’t seem to be a drastic difference in rates of signal detection by the qPCR instrument. I suppose it’s back to the drawing board for us, classic science! Don’t worry, we still have a couple more ideas.
As far as the manuscript writing process, I’m not as far as I may have ambitiously hoped to be at this point in the quarter. Though, all the hiccups with lab work and methods for the experiment have needed more attention than anticipated. I did, however, create a concept map and received feedback to draft an outline for my introduction. Some progress has been made!
Thank you!
Maddie Gard
Zoë's Blog
Zoë Lewis, graduate student
1 March 2022
With Spring around the corner, it’s starting to feel like the clock is ticking on my thesis deadlines! Although I’m still waiting on a few aspects of my data analysis, I finally feel like the DNA metabarcoding data is ready to be analyzed. Increased lab work and data analysis time has put me a bit behind on writing deadlines, however, my ambitious writing schedule set at the beginning of the year left plenty of time for hiccups. Although it is a bummer to miss self-set deadlines, the wiggle room in my schedule allows me to focus on the pressing research needs, and I think this flexibility has been crucial in the success of this project.
I mentioned in my last blog that I would hopefully have new data news to share, however, I’ve spent most of my time these past few weeks working on data cleaning and analysis, so I don’t have anything exciting to report! Given that many of these molecular techniques are accepted and verified for harbor seal scats, I’m digging deeper to ensure all these techniques will still hold for my dataset. This had involved changing some of the bioinformatics on the process, as well as looking at the impacts of improved DNA metabarcoding methods on DNA sequence reads returned.
Most critically, Maddie and I realized 3 weeks ago that the qPCR sex determination primer/probes designed for harbor seals do not work on Steller sea lion samples. I was not expecting this to be the case, based off our interrogations of known pinniped DNA, however after our preliminary results we found that this method was unreliable. So, this resulted in a large portion of my time spent on redesigning qPCR primer and probes for Steller sea lions (and learning how to design a qPCR primer and probe!), running preliminary tests and adapting run methods for the qPCR machine. I’m hopeful that this week will result in a method that works reliably for our samples… but stay tuned!
Kathleen's Blog
Kathleen McKeegan, graduate student
1 March 2022
February was a month full of presentations! On the 11th, I presented to the second-year fisheries students in Bellingham Technical College at Whatcom Creek Hatchery. I presented to their seminar class and focused on all the research that our lab has been conducting at the creek since 2011, including Grace Freeman’s thesis research and my research on TAST. On the 23rd, I presented at the Whatcom Marine Research Symposium hosted by the Whatcom Watersheds Information Network. The Symposium itself was well organized and included many fascinating presentations. I shared my research on TAST and how it is impacting the individual seals in Whatcom Creek. On the 28th, I presented my thesis research project to all the graduate students in the biology program for our Biology 523 class. I also had an informal outreach opportunity with a graduate student in the WWU English Department earlier this month. He reached out and we recorded a podcast in which I described my thesis research, the harbor seals at Whatcom Creek, and the implications of my study in how it relates to salmon recovery. It was a great experience and I look forward to hearing the final podcast soon!
Through these outreach opportunities, I have been able to improve my scientific communication skills and perfect the way I share my data. That said, I have spent a lot of time preparing for these various conferences and presentations, so I am feeling a little behind on my data analysis. I have completed much of the data visualization and summary statistics for 2020-only (comparing TAST on and off days), but Kate and I are still finishing the photo IDs for 2021. I am also working on the model fitting process for 2020 using a Generalized Linear Mixed Model with a Poisson log-link and ID as a random intercept. I am looking into possibly using a Zero-Inflated Poisson, but it may not be applicable to my dataset. I also just received the Chum salmon numbers from BTC so I can input those into the model and see how salmon numbers relate to seal success. I am still in the process of creating these statistical models, but from my early exploration, it does not seem like TAST is a significant predictor of individual foraging success. It will be interesting to see how this analysis progresses.
Besides presentations and analysis, our lab is still conducting weekly observations at the creek. We also just met with the instructors and researchers at BTC/Whatcom Creek Hatchery to discuss future collaborations between their students and our lab. There are a lot of exciting things in the works, but for now I am focused on finishing my data analysis, processing Fall 2021 photos, and writing my thesis. Furthermore, after much consideration, I have decided to push my graduation to summer quarter to allow for more time for data analysis and writing. Things are getting exciting, but I still have a lot to do! Hopefully I will have more to report next month!
Holland's Blog
Holland Conwell, undergraduate student
1 March 2022
Happy almost spring! Despite the currently gloomy sky, I have my heart set on some sunshine soon. Since last month, I had a bit of a setback while working in Excel, which forced me to take a couple steps backward. There was a mysterious loss of data from a couple locations, notably Belle Chain and Comox in 2012. I traced this back to an early error stemming from transforming Madelyn’s data from wide to long format with Power Query in Excel. It turns out that unpivoting in Power Query will automatically remove rows with no/null values, like what is found in my dataset…lesson learned.
Due to this computational betrayal, I decided to do as much of my work as possible in RStudio, starting with turning the data from wide to long format. I re-merged Madelyn Voelker’s data with mine from scratch, checking the Belle Chain and Comox 2012 data intermittently and tracking any and all progress that I made during this process. Once I got the overall merged dataset in wide format, I checked that the DNA diet fraction for all individual samples summed to 1 across all rows, with all but 26 samples summing to 1. That said, this appears to be a very small portion of samples, and a quick check revealed that these samples did not even sum to 1 in Madelyn’s original dataset. With the merged dataset, I created an updated plot of the distribution of sexes across locations, months, and years, and I was able to identify all locations lacking sexing information. I still need to clear up some confusion surrounding some samples, but I’m currently working on deleting sites from the dataset that have no sexing information in the plot, with the goal of condensing the dataset to only sites with both sex and diet data. With this condensed dataset, I will be able to start comparing sex to different variables and truly digging into data analysis!