This project is focused on revealing the earliest signs of motor decline in ALS by closely tracking movements in drosphilia models using tools like SLEAP. The goal is to improve our understanding of how the disease progresses and to pave a path for more effective treatments in the future.
Nidia Gutierrez, Oscar Osuna, Humna Rana, Walid El Boukhari, Andre Perez, Jocelynn Medina
It's a neurodegenerative disease that leads to the death of upper and lower motor neurons, disrupting signals to innervated muscles and resulting in muscle atrophy.
These motor neurons are essential for controlling voluntary muscle movements and respiration.
There are very few treatment options available, primarily limited to temporarily alleviate symptoms.
Molecularly, Drosophila are similar to humans, with a comparable genome consisting of approximately 14,000 genes, while humans have around 21,000.
These flies have a rapid reproductive cycle, which significantly reduces the time required for research.
Drosophila melanogaster share approximately 65% of human disease-associated genes, making them valuable for modeling human disorders.
This involved a cross between Gal4 virigin females and males from the GFP line.
This is a reporter gene expressed by GAL4 so degenerative effects aren't expected.
This involved a cross between Gal4 virgin females and SOD1 males.
This ALS gene consisted of two balancers, Tubby (Tb) and Stubble (Sb). We focused on the Tubby balancer, as it allowed us to distinguish the flies' genotype.
Cross between Gal4 virgins and male TDP-43
This ALS gene consisted of one balancer, Stubble (Sb)
Superoxide Dismutase 1 Gene (SOD1): Normally expresses an enzyme in the cytosol and mitochondria that clears out reactive oxygen species in the cell that otherwise can cause oxidative stress.
Mutations in SOD1 gene induces aggregation of the gene product, inducing dysfunction in mitochondria and facilitating death of upper and lower motor neurons.
TAR DNA Binding Protein 43 (TDP-43): normally expresses an RNA-binding protein that contributes to the formation of intracellular condensates that participate in RNA metabolism.
Mutations in this gene induce the formation of fibrillary tangles that are very difficult to clear out, causing neuronal dysfuntion and death.
Fly collection: We began by finding a labeled clear plastic vial filled with stock flies that contained the necessary genotype for the experiment. Once the correct vial was found we then used carbon dioxide in order to sedate the flies. As soon as they were sedated in the vial, the flies were subsequently placed on a porous pad that secreted more carbon dioxide in order to keep them sedated. Upon being placed on the pad, the team separated them into groups of females, virgin females, and males by using a microscope.
Crosses: The virgin females and males would be used for a genetic cross while any unused flies would be discarded in ethanol. The cross would then be closely monitored throughout a span of several days for pupil cases. The parents would then be flipped out of the cross vial after pupil cases began showing. Since both observed genes contained balancers it was vital to monitor the expression of the balancer phenotype such as a tubby SOD1 pupal case.
Behaviors: Once crosses are mature enough, two male offspring would be put into a tub of ice until lightly sedated (time varied between 2-5 minutes). When the two males were sedated, they were placed on a 3D printed arena. The arena was covered with a silicone petri dish that was coated with silicone in order to prevent the flies from climbing onto the roof and clouding motion tracking. The behavioral observation video was recorded with an iphone using 1080p and 60fps from 9 centimeters away and was recorded for approximately four minutes.
SLEAP: After the recording was finished, the two flies were sedated with carbon dioxide again through a small opening within the arena and placed back into their vials. The video is then uploaded to an AI software called SLEAP to analyze and quantify the flies' gait through motion tracking.
According to the ALS publication, the risk factors for ALS were slightly higher in males than in females, independent of age.
In our SOD1 line, we observed a low eclosure rate among the male offsprings and a higher rate among females, indicating potential male lethality prior to eclosure.
In our TDP-43 non-stubble offspring, we observed an approximately equal male-to-female ratio, with motor decline occurring within two weeks in males and slightly later in females by 21.
After recording videos, we uploaded them to the HPC user interface, and used the supervised machine learninSLEAP to:
(1) label the limbs, head, abdothorax, and wings of each fly in our recordings at 20 randomly chosen frames,
(2) train a model using those labeled frames,
(3) use the inference feature to generalize to all other frames of those videos, and assess the predictive power of the model we generated.
D1
FD1
FD28
FD14
One of our main limitations gathering data was not identifying balancers in the first month of the project
In our SOD1 line, there was no certainty that the collected offspring carried the SOD1 gene. Consequently, the video analysis from our initial SOD1 flies was considered invalid.
Another issue we encountered was the failure of reproductive success, which may have resulted from our overcrowded stock conditions or the age of our males affecting the reproductivity of our cross.
Video quality and setting variability gave us results of different videos notable difference when inputing into SLEAP program.
Wait time in between start of recording and sedating them in ice; affecting fly mobility in the arena causing erradic movement of the flies.
Issues with Interactive Desktop:
Getting an interactive desktop to work while there were several people working with them.
Frozen shrunk screen
Solved it through clicking every button. Next time, it should have been tracked how it was resolved incase the issue pops up again
Issues with SLEAP
Error when trying to run inference
It would run until 99% and then freeze for 10 minutes
It would run all the way through and then have an error pop up saying 0 inferences found
After resolving inference problem, the next step was uploading all of the videos
Tried doing it all at once
It shows on the bottom left corner that there are 1/44 videos in the project, however they were not showing in the video section
Markstein, M. (2018). Drosophila Workers Unite! A laboratory manual for working with Drosophila.
Prasad, A., Bharathi, V., Sivalingam, V., Girdhar, A., & Patel, B. K. (2019). Molecular Mechanisms of TDP-43 Misfolding and Pathology in Amyotrophic Lateral Sclerosis. Frontiers in Molecular Neuroscience, 12. https://doi.org/10.3389/fnmol.2019.00025
Ruhmann, H., Koppik, M., Wolfner, M. F., & Fricke, C. (2018). The impact of ageing on male reproductive success in Drosophila melanogaster. Experimental Gerontology, 103, 1–10. https://doi.org/10.1016/j.exger.2017.12.013
Sharad Kumar Suthar, & Lee, S.-Y. (2023). The Role of Superoxide Dismutase 1 in Amyotrophic Lateral Sclerosis: Identification of Signaling Pathways, Regulators, Molecular Interaction Networks, and Biological Functions through Bioinformatics. Brain Sciences, 13(1), 151–151. https://doi.org/10.3390/brainsci13010151