Melanin comes in two main forms.

Eumelanin is a black or brown pigment and constitutes about 74% of the melanin in human skin

Eumelanin causes darker skin, hair, and eyes and has photoprotective properties. UV-radiation can cause breakages in DNA strands which can lead to skin cancers such as basal cell carcinoma or melanoma. Additionally, UV radiation can create reactive oxygen species such as hydroperoxyl (HO₂) which can interfere with biological functions. Eumelanin can act as a physical barrier that absorbs and scatters UV radiation before it can reach the DNA. As a result, individuals with light skin are about seventy times more likely to develop skin cancer than individuals with darker skin.

Pheomelanin is a red or yellow pigment and is responsible for about 26% of melanin.

Pheomelanin, on the other hand, is a red and yellow pigment. In humans it can appear as red hair (blonde hair is caused by low overall melanin levels). Unlike eumelanin, pheomelanin is prone to degradation by UV light which can create hydrogen peroxide anions. This has negative effects on the skin and can result in increased cell death upon experiencing UV radiation. Pheomelanin is can cause increased histamine release which results in skin redness, especially in lighter skinned individuals.

Other forms of melanin include neuromelanin which colors the brain. Neuromelanin may be associated with neuroprotective effects as it is found in lower levels in Parkinson’s Disease patients. Additionally, neuromelanin is more prevalent in humans than any other primate.

Overall, melanin can cause a wide variety of effects in the body, including, but not limited to coloring our cells.

References

Nguyen, D. H. (n.d.). What colors are the cells in your body? Seattle PI. https://education.seattlepi.com/colors-cells-body-5632.html (Seattle PI - Education)

Nasti, T. H., & Timares, L. (2015). MC1R, eumelanin and pheomelanin: Their role in determining the susceptibility to skin cancer. Photochemistry and Photobiology, 91(1), 188–200. https://doi.org/10.1111/php.12335 (PMC)

Cleveland Clinic. (n.d.). Melanin: What is it, types & benefits. Cleveland Clinic. https://my.clevelandclinic.org/health/body/22615-melanin

D’Mello, S. A. N., Finlay, G. J., Baguley, B. C., & Askarian-Amiri, M. E. (2016). Signaling pathways in melanogenesis. Pigment Cell & Melanoma Research, 29(5), 528–540. https://doi.org/10.1111/pcmr.12410

Slominski, A., Tobin, D. J., Shibahara, S., & Wortsman, J. (2004). Melanin pigmentation in mammalian skin and its hormonal regulation. Physiological Reviews, 84(4), 1155–1228. https://doi.org/10.1152/physrev.00044.2003

Ito, S., & Wakamatsu, K. (2005). Chemistry of mixed melanogenesis—Pivotal roles of dopaquinone. Pigment Cell Research, 18(4), 264–272. https://doi.org/10.1111/j.1600-0749.2005.00244.x

Marie Curie

Photo from the Nobel Foundation Archive

Marie Curie was born in 1867 in Warsaw, Poland. Growing up Polish under the tsarist regime, her family was constantly in a precarious position, with her father needing to move from job to job. She attended Sorbonne University in Paris when she was 24 because the Russian Empire prohibited women from attending university. There, she did her thesis on radioactivity. Her research in radiation phenomena led to her earning the Nobel Prize in Physics in 1903 along with her husband Pierre Curie and French physicist Henri Becquerel. Her continued research in radiation led to her discoveries of polonium and radium. For these discoveries, she was awarded the 1911 Nobel Prize in Chemistry, making her both the first woman to earn a Nobel Prize and the only person to earn multiple Nobel Prizes in two different sciences.

Alice Augusta Ball

Alice Ball was born on July 24^th, 1892, in Seattle, Washington. She earned two degrees at University of Washington and was the first Black woman to graduate with a degree in pharmaceutical chemistry. After graduation she published a paper in the Journal of the American Chemical Society. She received a full scholarship for the College of Hawai’i. There, she studied methods to extract active ingredients from the kava root. This research was of interest to Dr. Harry T. Hollman and Dr. Arthur Dean who were studying cures for leprosy. They wanted her to replicate what she did with kava roots to chaulmoogra oil (the existing treatment). She extracted the active ingredient within the oil to create a treatment that the human body could easily absorb. This was the first leprosy treatment which halted disease progression. This treatment saved countless lives. Unfortunately, one year after her discovery, Alice Ball fell ill and died at the age of 24. Arthur Dean published her research as his own, giving her no credit, highlighting the discrimination she faced as a Black woman in a STEM field. Her legacy was restored in the 1970s when a Hawaiian historian named Kathryn Takara found her original thesis.

Inge Lehmann

Dr. Inge Lehmann (1888-1993), discoverer of the Earth’s inner core. Photo courtesy of B.A. Bolt.

Inge Lehmann was born in Denmark in 1888. She attended a progressive school that afforded women the same education as their male counterparts. In 1929, Lehmann studied the seismic waves from an earthquake in New Zealand. She found the waves were bouncing off of some kind of invisible boundary. She theorized that the Earth’s core was separated into a solid inner core and a liquid outer core. Her hypothesis was confirmed in 1970 when more sensitive seismographs better detected waves deflecting off the solid inner core.

Rebecca Lee Crumpler

*there are no surviving pictures of Rebecca Lee Crumpler

Rebecca Lee Crumpler was born in Delaware in 1831. She grew up with her aunt who frequently cared for their sick neighbors. This inspired her to enter the medical field. In the 1850s, she moved to Massachusetts to pursue her career as a nurse. She assisted various doctors across Boston. In 1860, she became the first African American woman to enroll in the New England Female Medical College in Boston. The college was one of the only in the country that awarded medical degrees to women and allowed Black students to enroll. On March 1, 1864, she received a Doctress of Medicine at the age of 33. She was the first Black woman in the United States to receive a medical degree. After the Civil War ended, she prioritized helping newly emancipated slaves who had little access to healthcare. Despite facing extreme racism and sexism, she did not stop her work as a doctor. In 1883 she published a book called A Book of Medical Discourses based on journal notes from her years as a physician. This book is widely thought to be the first medical text written by an African American author.

These women are a reminder of how breaking barriers and improving access can result in more innovation.

Olbers’s Paradox:

Olbers’s paradox was popularized by Heinrich Olbers (though the theory can be traced back to 1601). It highlighted problems with the model of a static and eternal universe before this was proved formally. Essentially, the paradox states that if the universe were static and timeless, there would be an infinite number of stars. The light from these stars also would have had time to travel towards us. Thus, the night sky would appear completely white due to the light from these stars. Since this is not true, the universe cannot be static and timeless.

The Maggot Milkshake:
Entomotoxicology is a branch of forensic science that uses bugs found on deceased bodies to glean information about the cause of death. When forensic scientists find a decaying body, they need to test it for toxins to determine the cause of death. Decaying bodies usually come with a host of maggots eating the body. If the deceased had ingested any drugs prior to death, those drugs will be in the maggots. So, researchers blend them into a “Maggot Milkshake” that they use to test for toxins. This is especially helpful for cases where the body has decomposed to the point where testing becomes difficult.

Finger Muscles:

Did you know that you do not have muscles in your fingers? Instead, finger movement occurs due to muscles in the forearm with finer movements controlled by small muscles in the hand.

Cherenkov Radiation

The characteristic blue glow of a nuclear reactor (pictured above) is caused by a phenomenon similar to that of a sonic boom. A sonic boom occurs when an object travels through air faster than the speed of sound. Cherenkov radiation is the same principle, but with light in a medium. In a clear medium such as water, it is possible for particles to move faster than light because in water, light is slowed to only 75% of its original speed. In a nuclear reactor, the charged particles released by the reactor move faster than light. These particles release photons (particles of light). These photons have a high frequency and a short wavelength, so the light appears blue to the naked eye.

Early phylogeny was based on morphological or physical differences between animals which led to many classification errors. For example, shared morphological similarities between pangolins and armadillos such as their long snouts, ant-eating diet, and armored scale-like structures led them to be classified together in the order Edentata. In reality, pangolins are the last living members of the order Pholidota, and armadillos belong to the order Xenarthra.

Armadillo:

Pangolin:

As the field of genetics was discovered and the advent of computers improved data processing capabilities, phylogenetic trees were increasingly based on cladistics: the field of study that classifies evolutionary lineages based on characteristics shared by a common ancestor and all of its descendants. This classification is used to arrange evolutionary lineages into monophyletic groups called clades. Because modern cladistic studies largely rely on molecular data, it yields more accurate results when it comes to decoding evolutionary ancestry. Relying on monophyletic groups for classification also makes it easier to track evolutionary divergence.

Understanding evolutionary groups as clades leads to some interesting classifications that may not occur to you upon first glance. For example, modern birds are descended from dinosaurs. To create a clade that includes all of Dinosauria, one would have to include birds. Thus, birds are dinosaurs.

This logic can be applied to other terms as well. The word fish, for example, colloquially refers to aquatic vertebrates. This is a paraphyletic group that includes jawless fish, cartilaginous fish, and bony fish. The problem is that paraphyletic groupings are based on arbitrary perceived differences. Thus, they don’t capture the full picture of evolutionary lineages and lose value as a classification tool. From a purely cladistics perspective, because tetrapods (and thus humans) evolved from lobe-finned fish, humans must also be included in the classification of fish.

How Do Placebos Work?

The placebo response is a result of a patient’s expectations and social cognition. Despite having no effect to treat the patient’s illness, placebos are administered in a medical context. Patients have an expectation that they will get help because of this context. This elicits a response from the brain that decreases pain and stress (Fun fact: a placebo that elicits a negative response is called a nocebo). Placebo analgesia (reduced pain) is associated with the autonomic nervous system and the endocrine system, indicating that the placebo effect has physiological pathways. A 2013 study by Stephan Geuter and Christian Büchel found that nocebo treatments can activate pain-facilitating mechanisms in the spinal cord. Verbal nocebo treatments have been associated with hyperactivity in the neuro-endocrine system and increased levels of cortisol and adrenocorticotropic hormone. This all points to the fact that the placebo response is not purely psychological and is also rooted in neurobiology.

Placebos as a Potential Treatment

While placebos do not work to cure an underlying condition, they do help with somatic symptoms such as pain and nausea. In fact, placebo treatments have been recorded to improve motor skills in Parkinson’s disease patients due to increased dopamine release. Additionally, the difference between placebos and antidepressants is modest when looking at mild to moderate depression. This is important because it means that placebos could potentially serve as a supplement to drugs.

Does knowledge of a placebo impact its effect?

One barrier to the use of placebos as a treatment is patient-provider transparency. Since placebos seemingly work due to the expectation that the treatment will improve symptoms, it stands to reason that knowingly taking a placebo would eliminate its effectiveness. This poses ethical concerns as if a physician tells a patient that a placebo is an active drug when it actually isn’t, there is a breach of trust. Interestingly however, the placebo response holds even when a patient is aware that they are taking a placebo. A 2021 study by Lembo et al. compared the progression of Irritable Bowel Symptoms across three groups: no pill, open label placebo, and a double-blind placebo.

What they found was that the group that knew they were taking a placebo still showed improvement in symptoms compared to the group that took no pill, as seen in the above figure. Open label placebos have also been shown to improve outcomes in opioid use disorder and chronic lower back pain. While open label placebos may not be as effective as double-blind placebos, the lack of deception makes them a promising method to alleviate patient symptoms.

Ultimately, placebos are a testament to the versatility of the human brain. It is important to continue researching both placebo and nocebo responses because understanding how our brains respond to placebos can help improve understanding of disease pathways and how to treat them.

The crow has demonstrated facial recognition skills. In 2006, University of Washington professor John Marzluff and his students wore a mask, dubbed the “dangerous face.” Crows were exposed to this mask as the team netted and trapped between 7-15 birds at multiple different sites in Seattle, WA. What they observed was that prior to trapping, the crows had little reaction to people wearing the mask, but after trapping, crows would attack and scold people wearing the mask. Additionally, they did not express this hostile behavior to people not wearing the mask. This indicates that the crows accurately identified the faces of those that trapped them and had not developed a newfound hostility towards all people.

But why does it matter that crows can recognize your face? Facial recognition is demonstrative of greater memory abilities within crows. This, paired with crows’ high adaptability when it comes to tool use, shows that crows are not passive observers of the world with visual cues leaving their brains as fast as they enter.

Another bird that demonstrates high intelligence is the parrot. Animal behavioralist Irene Pepperberg studied Alex the African grey parrot for 30 years. Alex is the first nonhuman animal to have ever asked a question. Specifically, he asked “what color?” upon seeing himself in the mirror. This shows that Alex was not just repeating words he heard. He was able to use language to acquire new knowledge. Since this question was in response to his own reflection, it also shows that Alex had self-awareness. Over the 30 years Pepperberg studied him, Alex demonstrated many other cognitive abilities such as human-level understanding of the Müller-Lyer illusion and understanding the concept of zero. In doing so, Alex challenges the idea that human cognition is uniquely superior.

The intelligence of crows and parrots serves as a reminder that nonhuman animals are not mindless creatures floating through life. Internalizing that these animals are as aware as us can help us remember to always respect the world we are a part of.

This “othering” of the environment likely stems from humankind’s idea that humans are too smart to be an animal. Humans cook their food and do arithmetic and build cities. Every species has unique characteristics, though. Being unique doesn’t remove humans from the animal kingdom. When it comes to intelligence, though, we tend to neglect the intelligence of other animals.

A 2019 study by Howard et al. demonstrated that honeybees were taught and understood basic arithmetic. Bees are tiny, and their brains are tinier. But they can add. This example is notable to me because when I try to think of nonhuman animals that are “intelligent,” my brain jumps to apes or dolphins or crows. Animals that are not that far from us, evolutionarily speaking. But bees? Bees are invertebrates, so their lineage diverged from ours hundreds of millions of years ago. Despite so many millions of years separating our species, similarities in our cognition persist. Another example of an invertebrate similar to humans is the octopus.

The nervous system of an octopus is extremely decentralized. While they do have a brain in their head, each of their arms can act like an individual unit, making an octopus seem like nine connected creatures. That being said, octopuses do exhibit similar behaviors to humans. The body transfer illusion is an experiment in which a person’s hand is hidden, and a rubber hand is placed in front of them. Then, both their hand and the rubber hand are prodded in the same way. Then, the rubber hand was struck with a hammer. The experiment showed that people would react as if their own hand was struck. Recently, in a study by Sumire Kawashima and Yuzuru Ikeda, this same phenomenon was demonstrated in octopuses. Despite our differences in physical structure, our minds still function in similar ways.

Humans aren’t the exact same as every other animal, but we aren’t completely different either. That’s what it means to be a species. Earth is home to us all and internalizing that fact can help remind us that hurting the environment hurts us too.

The TME refers to the area around and including a tumor. This microenvironment includes tumor cells, non-cancerous cells, blood vessels, stromal cells, immune cells, and extracellular macromolecules. This environment lacks oxygen and is extremely acidic, so the TME promotes blood vessel creation to prevent tumor death and immune cells in the TME both enhance and fight tumorigenesis.

The hypoxic and acidic nature of the TME means that tumors constantly need to fight for resources. The researchers extracted human white adipose tissue (WAT) for in vitro studies because these cells are easily extracted via liposuction. Then, UCP1-CRISPR and doxycycline (which allows for reversable gene activation) was used to activate these cells, making them more similar to beige adipose tissue which can convert energy to heat. This increases the metabolic rate of the WAT tissue, allowing it to suppress tumors via competition for glucose and fatty acids. These modified cells were observed with breast, colon, pancreatic, and prostate cancer cells and the modified WAT tissue was shown to suppress cancer proliferation and reduce the tumor’s ability for glucose uptake. The below image shows the difference in tumor size between a control tumor and tumors injected with UCP1-CRISPR WAT tissue.

The modified WAT tissue was then implanted into breast cancer and pancreatic cancer mouse models, and the tumor suppressive properties were maintained. The breast cancer mouse model also demonstrated that the WAT tissue could be placed at the center of the tumor and farther from the tumor with no effect on the results. This distal placement of the WAT tissue implies that if CRISPR modified white adipose tissue becomes a viable cancer treatment option, implantation will not necessarily depend on how accessible the tumor is via surgery.

It is important to note that more research must be done into how the WAT cells interact with healthy cells around the tumor. Although doxycycline allows for a reversable treatment, future studies must address long term effects before this could become a treatment option. Still, adipocyte implantation proposes a method of fighting tumors indirectly through the TME and will hopefully become more promising as more research is done.

Works Cited

Anderson, Nicole M, and M Celeste Simon. “The tumor microenvironment.” Current biology : CB vol. 30,16 (2020): R921-R925. doi:10.1016/j.cub.2020.06.081

Nguyen, H.P., An, K., Ito, Y. et al. Implantation of engineered adipocytes suppresses tumor progression in cancer models. Nat Biotechnol 43, 1979–1995 (2025). https://doi.org/10.1038/s41587-024-02551-2

As it turns out, plants can suffer from tumors, but not exactly what we would call cancer. For example, the bacterium Agrobacterium tumefaciens can cause uncontrolled cell division resulting in bulges called crown galls.

Interestingly, tumors are less deadly for plants than they are for animals. This is because plant cells have a cell wall. The cell wall is fairly rigid and allows for strong adhesions between cells. Even if a tumor does form, the cell wall prevents it from migrating. Thus, even though plants can get tumors, they rarely get malignant ones.

Additionally, in humans, cancer often spreads by traveling through the bloodstream. Plants don’t have a circulatory system like our bloodstream and instead use fluids such as sap to spread nutrients. Since sap does not have cells, tumor growths in plants cannot easily spread.

Studying how plant cells prevent cancer can help to improve understanding of cancer in humans. While we don’t have a cell wall, a 2023 paper by Thibault Leroy supports the conclusion that trees are more likely to experience mutations as they age rather than from increased cell division. Understanding how plants protect against mutations can allow scientists to utilize those methods in humans.

This observation has led scientists to conclude that larger animals likely have better cell cycle regulation than smaller ones. Thus, by studying the cells of a whale, for example, scientists could gain insight in how to prevent cancer in humans. That is exactly what Firsanov et al. did in their recent study Evidence for improved DNA repair in the long-lived bowhead whale. Despite being the second largest animal on Earth, the bowhead whale does not see high incidence of cancer. When comparing human and bowhead whale cells, they found that the whale cells exhibited a higher ability to faithfully repair double strand breaks in DNA, indicating lower mutation rate overall.

The researchers identified this to likely be due to the cold-inducible RNA binding protein (CIRBP) which is associated with cell repair in humans. CIRBP is highly expressed in the bowhead whales’ fibroblasts and tissues. When the bowhead CIRPB was introduced to human cells, the cells exhibited increased homologous recombination and DNA end protection. In Drosophila, introduction of CIRPB increased both the insects’ lifespans and their resistance to ionizing radiation.

These results show promise in the development of cancer prevention drugs via DNA repair. This could help the elderly and people who are genetically predisposed for cancer and possibly result in lower cancer rates in populations as a whole. Of course, more research needs to be done on the safety and viability of these treatments, but this development is very exciting news in the long fight against cancer.

Alzheimer’s disease is a chronic disease characterized by neurodegeneration and the buildup of amyloid proteins. It is also the 7^th leading cause of death in the United States. Right now, 10% of people over 65 and it is estimated that over 42% of people over the age of 55 will develop dementia at some point. This means that Alzheimer’s disease is on its way to become a larger issue, so it is important to investigate ways to prevent Alzheimer’s disease.

Today I am going to be explaining how vitamin D supplementation could help prevent the onset of Alzheimer’s disease in people who are vitamin D deficient.

Vitamin D is a fat-soluble vitamin that is usually obtained naturally in the human body through exposure to the sun and is found in a few foods such as fish and eggs. Over 35% of the United States population is vitamin D deficient, making it a widespread issue. According to the NIH, vitamin D improves calcium absorbance in the body, making it commonly associated with bone health, but there are vitamin D receptors in the brain.

In the brain, vitamin D improves the expression of LRP-1. This is important because LRP-1 facilitates the movement of Amyloid-Beta proteins out of the brain. When LRP-1 expression is reduced (which could happen in a vitamin D deficient person), the Amyloid-Beta proteins can buildup in the brain. This is one of the things that leads to Alzheimer’s onset. Thus, vitamin D supplementation in deficient people could decrease buildup of Amyloid Beta proteins, decreasing rates of Alzheimer’s development.

Above is calcitriol which is the form of vitamin D active in the previous diagram and is a steroid hormone and it’s the active form of vitamin D in the body. It finds to the vitamin D receptor that is found in the cell nucleus, and it is produced in the kidneys through the hydroxylation of calcidiol.

Calcidiol (one of the names for the storage form of vitamin D) is produced in the liver through the hydroxylation of vitamin D₃, and this is what you measure to determine vitamin D deficiency.

The first study that I looked at is a meta-analysis of many different studies that shows that vitamin D deficiency is associated with increased risk of Alzheimer’s.

The researchers eliminated studies from their pool based on baseline dementia, the threshold for vitamin D deficiency (the paper considers less than 75 ng/mL to be deficient), and a lack of sufficient data. 23 papers were ultimately identified, and the researchers ran Random Effect models to determine the level at which vitamin D deficiency is linked to the onset of dementia, Alzheimer’s, and cognitive impairment.

I focused on the result of vitamin D deficiency on Alzheimer’s patients. On the left are the different studies and on the right are the risk ratios and confidence intervals of said ratios. The random effects model found a risk ratio of 1.57. This means that, overall, people with vitamin D deficiency are about 50% more likely to develop Alzheimer’s than someone who is not deficient.

Since the meta-analysis demonstrates a correlation between vitamin D deficiency and Alzheimer’s onset, I found this animal study which demonstrates the helpful effects of vitamin D supplementation in providing neuroprotection against Alzheimer’s.

The researchers created an induced Alzheimer’s model in rats by feeding them aluminum-chloride-D-galactose (AlCl3-D-gal) which induced mitochondrial dysfunction. These mimicked effects of Alzheimer’s because mitochondrial dysfunction is a symptom associated with Alzheimer’s disease. The rats were split into four groups based on diet (a control, a vitamin D rich diet, the Alzheimer’s model diet, and a group fed both the vitamin D rich diet and the Alzheimer’s modeling one) and fed for 51 days. Then ten days were spent on behavioral trials.

The first graph is from a Novel Object Recognition test. The control, vitamin D groups along with the group that was fed both Alzheimer’s and vitamin D diets show much higher recognition of the old object compared to the rat group only fed the Alzheimer’s diet. This indicates that the vitamin D is providing neuroprotection against Alzheimer’s since the group being fed both diets shows higher memory levels than the Alzheimer’s only model. The second graph shows the results from a Morris Water Maze and the rats only fed the Alzheimer’s take much longer to find the escape platform than every other group which, once again, indicates that the vitamin D is providing neuroprotection since the group being fed both diets was able to find the platform about as quickly as the control group was.

Then the researchers excised the rat brains to observe changes on the molecular level. The group only being fed the Alzheimer’s inducing diet shows higher levels of the amyloid beta proteins than every other group, indicating that the vitamin D is working to move the Amyloid beta proteins out of the brains in the rats fed vitamin D along with the Alzheimer’s inducing diet.

Once I found an animal model that supported the use of vitamin D in preventing Alzheimer’s, I wanted a study that looked into what happened in humans. What I found was the following study: Vitamin D Supplementation Worsens Alzheimer’s Progression: Animal Model and Human Cohort Studies.

The researchers identified people from the National Health Insurance Research Database who were over 65 and dementia free before the year 2000 and assigned them to a treatment or control group. The treatment group had been prescribed calcitriol between 2000 and 2009, and the rate of dementia development was observed over 10 years.

This graph indicates that people taking over 10.95 micrograms instantly have increased rates of dementia. The thing is, this graph looks sketchy, since the change in dementia rates happens instantly. With something such as supplementation, the results are expected to take some time to show up. The instantaneous drop off indicates that the rates of dementia are caused by something other than the higher vitamin D dosage. For example, this study did not look at why these supplements were being prescribed, and people who are being prescribed vitamin D are the same people who are already at higher risk of Alzheimer’s due to poor health.

Because of that study’s questionable results, I found another human cohort study that did take into account backgrounds such as race, gender, and prior medical history when looking at the effects of vitamin D.

From the larger dataset, people were eliminated based on missing follow-up visits and already having dementia at the baseline. Then the study participants were split up into two groups, those taking no vitamin D supplements and those who were.

Here the blue line represents the people in the study exposed to vitamin D while the green line represents the people in the study not exposed to vitamin D. The group not exposed to vitamin D shows decreased dementia survival probability than the group exposed to vitamin D and this gap widens over time. The thing is, here this change is not instant, and the lines are aligned for about a year into the study. This provides kore confidence that the results are due to vitamin D supplementation rather than some external factor.

Overall, my studies show that vitamin D deficiency is associated with increased rates of Alzheimer’s and that vitamin D supplementation is associated with decreased rates of Alzheimer’s development. While my third study did have results indicating otherwise, this study also had issues.

Specifically, Study 3 did not account for the backgrounds of patients or why they were prescribed vitamin D. Since people prescribed vitamin D are already at higher risk of Alzheimer’s due to poor health, not accounting for that could have caused the differing results. Also, the other two papers looking at humans (study 1 and study 4) both consider prior health conditions and backgrounds when presenting their data.

In the future, researchers looking more into the role of vitamin D in Alzheimer's prevention could investigate whether the type of vitamin D (like calcitriol or calcidiol) impacts the extent of Alzheimer’s prevention. Additionally, one could look at whether the dosage of vitamin D impacts the level of neuroprotection provided. Study 3 did look at this, but since it had issues, it would be interesting to read a study that looked at dosage while still accounting for the patient’s backgrounds.

Thank you so much for reading!

The topic that I have chosen to study is the effect of Vitamin D supplementation on Alzheimer’s Disease. When we first introduced our topics, the doctor in charge of the internship gave everyone feedback about directions to go in. My feedback was that while my topic was interesting, I needed to make sure that what I was presenting wasn’t a “duh statement.” This is because Vitamin D is generally accepted to have positive effects on the brain, and me presenting on the beneficial effects of Vitamin D could seem unsubstantial. Last Thursday, I presented my first article. Keeping the feedback in mind, I specifically presented an article that showed that Vitamin D did not have a positive effect on Alzheimer’s Disease.

The article that I presented was called Vitamin D Supplementation Worsens Alzheimer's Progression: Animal Model and Human Cohort Studies by Lai et. al. To start off my presentation, I gave some background on Vitamin D and why it is important to look into possible Alzheimer’s treatments. Vitamin D is a fat-soluble vitamin obtained through sun exposure and foods such as fish and eggs. According to the NIH, vitamin D is associated with calcium absorbance in the gut and is important for bone health, and Vitamin D receptors are found in glial cells in the brain. This means that Vitamin D does have some impact on brain function. Alzheimer’s Disease is an increasing concern in the health of the elderly. Currently, 10% of people over the age of 65 and have Alzheimer’s, and according to an NIH article from February 2025, it is estimated that 42% of adults over 55 will develop dementia. On top of this, current medications used to treat Alzheimer’s target symptoms of the disease but do little to actually stop disease progression. Also, earlier observational studies have led people to believe that Vitamin D supplementation will provide neuroprotection against Alzheimer’s which is why it is important to do more tests to confirm this. Even more so since Vitamin D’s impact in the elderly is poorly understood.

In the study that I presented, the researchers measured two forms of Vitamin D: Calcitriol and Calcidiol. Calcitriol is a steroid hormone and the active form of vitamin D. It is mde in the kidneys and binds to vitamin D receptor in cell nucleus. Calcidiol is the storage form of vitamin D and is produced in the liver via the hydroxylation of Vitamin D. This is what is measured to determine Vitamin D deficiency.

Subscribe now

The first part of the study was an animal trial. The researchers used both wild type mice and a special type of mice called APP/PS1 mice. These mice were genetically modified to express a protein that is commonly associated in human Alzheimer’s patients to induce an Alzheimer’s like state on the mice. Both sets of mice were then fed a Vitamin D rich diet and observed in two-month intervals. The results showed that both four months and six months after the start of observation, the vitamin D levels in the APP/PS1 mice were significantly lower than that in the wild type mice. This could indicate that Alzheimer’s causes vitamin D deficiency. Additionally, cell cultures were taken from the APP/PS1 mice and exposed to Amyloid Beta proteins (a type of protein associated with the brains of Alzheimer’s patients) and either calcitriol or calcidiol. In both cases, there were significantly less viable cells as the dosage increased, indicating that calcitriol and/or calcidiol supplements may have a negative effect on Alzheimer’s patients.

The next part of the study was two different human trials. The first used data from the National Health Insurance Database and the Longitudinal Health Insurance Database 2000 to observe the rates of developing Alzheimer’s for people over the age of 65 not prescribed calcitriol supplements compared to people who were prescribed calcitriol. The results indicated that people who did not take calcitriol supplements and the people prescribed the smallest doses of these supplements had relatively similar rates of developing dementia. Those who took supplements in doses of over 10.95 micrograms, however, were much more likely of developing dementia.

The third part of this study had a sample size of 980 people with dementia. 196 of them had a calcitriol prescription after developing dementia and the other 784 did not use calcitriol. The people were put in groups of one calcitriol user and four nonusers and groups accounted for age and gender. Interestingly, after being diagnosed with dementia, people that took the lowest dosage of calcitriol had a higher survival rate than people that took no supplements (as opposed to having the same rate of developing dementia). Still, those who took supplements of over 10.95 micrograms had a much lower survival rate.

After presenting the methods of the study, I went on to describe the possible limitations of it. Firstly, the study does not account for prior medical history, which means that it is impossible to factor in other causes for death. When I presented this to the group, the doctor in charge pointed out that people who are prescribed Vitamin D are usually getting such supplements because they have poor health already and are thus already at a higher risk for developing dementia. Also, the animal studies were done on mice which have very different brains from humans which means that the results from this study cannot necessarily be generalized to us.

Going forward, I want to research more about how Alzheimer’s progresses as a disease and how Vitamin D affects the brain in general. This way I will be more informed when I give my final presentation and be able to better understand the mechanisms behind the processes of Vitamin D impacting Alzheimer’s. And, of course, I need to read more articles about Alzheimer’s and Vitamin D to increase my understanding of the topic.

Going into this program, I was expecting to see and interact with doctors. Over the course of the program, however, I saw maybe six doctors in total. And when I did see them, they would go in to see a patient and then disappear. I knew on some level that nurses do most of the work at a hospital, but I did not realize that doctors would be an almost nonexistent presence. Instead, I mainly helped nurses with their daily tasks. Most of the work I did pertained to taking calls from patients and relaying requests back to nurses. I also did a lot of restocking storage equipment and replacing trash. I was always ready to take vitals for a patient, take samples down to the lab, or take a patient down for discharge.

The thing with the later tasks is that I had to take initiative to ask for work to do. The biggest advice about these programs that I have is that the more enthusiastic you are about doing tasks, the more tasks they give you. Obviously, the amount of work to do was tied to how busy the hospital was, but even on relatively empty days there is still a lot to do. The more work that I did without being asked (like restocking the fridge as soon as I arrived), the more tasks I was asked to do.

First, I tested the enrichment of the two top clusters of transcription factors (module1 and module2). The enrichment of a cluster means cross checking the genes in the module with genes that are known to be targets of known “important” transcription factors such as TP53 and BRCA1. For example, TP53 has 164 main target genes and 11 of them appear in module1 (as shown below). The p-value, 2.108368e-08 is extremely small, indicating that this overlap is of statistical significance and is not due to chance. Similarly, in module2, 9 out of the 57 targets of BRCA1 are present, with a p-value of 6.462656e-09. In fact, out of the top ten transcription factors cross checked with my list, each had an extremely small p-value—especially in module2. This all indicates that the genes that I have identified are highly likely to be indicators of cancer gene expression and are not a fluke.

After this, I wanted to test how many of the genetic relationships that I graphed last week are already recorded. I again used the TRRUST database, but this time I used their downloadable file of all 8,427 regulatory links between 795 transcription factors and 2,067 non transcription factors in the human genome. I figured that if it is not in that list, it would be safe to assume that I found something new.

In my analysis, I found 8,637 regulatory links. This is already more than the number in the TRRUSR database. When I compared my findings to those in the TRRUST database, I found only 26 overlaps. This leaves 8,611 possible regulatory links for me to explore.

Since I had so many possibilities for matches, I wanted to narrow it down. I did this by sorting the connections between transcription factors and target genes by Lasso coefficient. The larger the magnitude (or absolute value) of the Lasso coefficient, the greater the chance that the connection is statistically significant. I also removed all of the instances of self-regulation so that I am only looking at inter-gene interactions.

The issue now is that I need to find out what these connections mean—if anything. The way to do this would be to see if there is a pattern in the way these are expressed in patients and the patients’ outcomes with breast cancer. The issue is that patient data from medical studies is usually classified due to patient privacy and protection laws like HIPAA, so I need to brainstorm what I can do with my findings.

Thank you so much for reading and please stay tuned for updates!

Then I made the graph by adding edges to connect the transcription factor nodes. I gave the nodes weight so that they would appear in color. Then I used matplotlib and the network drawing tools to create the graph. I color coded the edges with a cold to warm gradient with blue corresponding to the lowest values and red corresponding to the highest ones.

The graph above has red loops around it. These loops indicate self-regulation. Since I wanted to focus on cross gene interactions, I got rid of the loops and remade the graph. I did this by saving a copy (so that I would not lose the original graph) and removing edges that pointed back on the original node.

Then, I ran some tests to find which transcription factors are the most central. First, I used the .out_degree() and .in_degree() actions to count which transcription factors regulate the most other genes and which transcription factors are regulated by the most genes respectively. The number TF hub, regulating the most other genes, is FOSL1 and the number one regulated gene is ALOX15B.

Then I tested for betweenness and eigenvector centrality. Betweenness centrality is when a node is in the middle of a path. This would act as a bridge. When I ran this on the graph data, I got a bunch of zeros. This means that the transcription factor network path lengths are all one. Eigenvector centrality denotes how important a node is based on how many hubs it is neighbors to. So if a transcription factor is not a hub, but it is connected to one, it would be denoted as more important. The results of this are pictured below.

I then used the Girvan-Newman algorithm to determine if there were any communities of transcription factor interaction. Even though each path was very short and no transcription factor lay in the middle of a path, there are two big communities of transcription factor interaction. The first has 108 nodes and the second has 107 nodes. The other hubs were all small, with 4 nodes or less. I then checked which of the ten top transcription factors that I had found are in which community, as shown below.

The two larger hubs are what my future analysis will most likely focus on, as they indicate a network of tightly connected transcription factors. Still, I do have to then cross-check this analysis with an outside source such as the TRRUST database which holds data of human transcription factor networks (which is what I plan to do in the coming week).

Thank you so much for reading and please stay tuned for more updates!

The data set that I used comes from the University of Santa Cruz and has an extensive list of gene expressions in 1218 patients. I checked that a few important transcription factors were present in the data and got confirmation, so I continued with my analysis.

Then I recreated my data sets of the transcription factors and cancer genes. Then, I reran the Lasso code on TP53. What I found was that, when in a data set with countless other genes, the patterns found are that TP53 is largely self-regulated.

I then checked a few more genes to see how many transcription factors regulated their expression. In doing this, I found that many genes are largely self-regulated. I did find that CCND1 expression, however, is regulated by numerous genes.

Then, I ran the same test I did for TP53 to find the transcription network of CCND1. The data (pictured below) means that the computer software found patterns in the data indicating a connection between CCND1 and the twenty other genes. This could indicate a transcriptional network of genes that regulate CCND1 but is not a 100% guarantee.

Finally, I used a loop to do the same analysis for all genes in the list of breast cancer related genes. The reason over 8000 rows is included is because each transcription factor could be involved in the regulation of multiple genes (so it is not just a list of each transcription factor). A snapshot of the table is pictured below.

In the next weeks, there are two avenues I plan to pursue.

First, I plan to use the networkx library in Python to create a visual model of my data. This will allow me to actually see what I am trying to analyze, which is helpful because only looking at the numbers makes it hard to see patterns due to the sheer amount of data. For example, it is virtually impossible for me to look at over 8000 data points in a timely manner, but a physical representation would allow me to see transcription factors that are hubs for interactions.

Second, I plan to use existing transcriptional regulatory network data to compare with my findings. It will give me a way to fact check what I have found. Existing data that matches my findings corroborates my analysis. Differences either mean that there is some discrepancy in my analysis or that I have found something new.

Thank you so much for reading!

First, I downloaded data of known cancer genes and transcription factors. I then used the DataFrame.columns function in pandas to create a list including the Hugo Symbols (identifiers) for genes found in the experimental breast cancer dataset.

The next step was to find matches between the experimental data and the known data. Using the intersection function, I found the intersections between the experimental data set with the cancer genes and transcription factors respectively. In doing this, I found 119 matches out of 249 cancer genes and 1020 matches out of 1640 transcription factors.

Then, I used the matching genes to splice the original transcription factor data and save it under new lists. One list (tf_expr in the first code snippet below) was the experimental expression of the matching known transcription factors while the other list (cancer_expr in the second code snipped below) was the experimental expression of known cancer related genes.

I then used the Lasso model in the scikit library to find correlations between transcription factors. The Lasso model is a statistical model that analyzes data to identify important values and predicts an outcome. In this case, the model predicted expression of a certain gene and which transcription factors were the most connected to the gene’s expression. Essentially, I created a basic model to predict networks of transcription factors. For example, I mapped 30 genes that were predicted to be connected to TP53: And TP53 is just one example. I was also able to map other genes such as MYC and VDR.

This is where the trouble started, though. After mapping those three genes, I tried mapping BRCA1. That is when I realized that my experimental data set did not have BRCA1 data. This was not the only missing gene. Other breast cancer related genes such as FOXA1, ESR1, GATA3, and PALB2, to name a few, were also missing. This is when I realized that the experimental dataset that I have been using has been filtered of some key protein coding regions. This means that I am missing out on some important genetic networks.

I have found a more comprehensive genetic network, though, so next week I will redo the work I have already done with the new data. This should hopefully be done faster since I have already done it once and my school workload is reducing as the end of the school year approaches.

After I am done redoing the progress I have made, I will use Lasso to analyze all known matching cancer genes in the list. Then I will compare the transcription factor networks I find to known networks. This will tell me whether what I have found is known or new.

Thank you so much for reading and please stay tuned for future updates!

First, I watched a four-hour long course on data analysis using python. This course outlined python libraries such as NumPy and Pandas. The NumPy library supports the storage of large arrays and matrices which can be used to efficiently store numbers. The Pandas library also allows storage in a manner similar to that of arrays. The storage structure used in Pandas is called a Series. When storing in an array, the stored items are ordered by an index. These induces are integers starting from zero. In a Series, a user can either use the integer indices but can also manually input them. For example, if a list of the number of mutations in a gene is stored by the Series mutationnum = [57, 12, 34, 28] then the user can either use the numerical index for each number (0 marks 57 mutations, 1 marks 12 mutations, 2 marks 34 mutations, and 3 marks 28 mutations), or the user can manually input index names so that each mutation number can be called using the names of the gene.

After watching the Python course, I set to work downloading data from the Cancer Genome Atlas BRCA data collection experiment. I downloaded two different files, one containing patient information and another containing information about mRNA expressions within the patients. I fed these into Jupyter as Series, an online python compiler, and did some preliminary analysis to practice using Java. Using the .info() function, I saw that the patient information file had 1101 rows, meaning that there are 1099 patients in the study, since the first two rows are table headers. This file also had 110 columns, meaning that there were 109 columns of different facts about each patient. The mRNA file again showed that there were 1100 patients. This file also showed that there are 20,530 different genes that were found to be mutated. Using the .value_counts() function on the patient information file on the column about a prior diagnosis, it was found that of the patients for which this information is available, 1028 have not had a prior cancer diagnosis and 68 have.

The next order of business is probably matching the gene IDs in the mRNA expression data to the gene IDs of known transcription factor coding regions. To do this I will first need to find a list of these known transcription factors and then download them so that I can use the coding software to compare the gene IDs. This will allow me to compare mutation numbers between genes that do and do not code for transcription factors.

Thank you so much for reading and please stay tuned for future updates!

This week, I read a bunch more articles about genetic regulation via transcription factors in the context of cancer:

• ARACNE: An Algorithm for the Reconstruction of Gene Regulatory Networks in a Mammalian Cellular Context by Margolin et. al

• Functional Characterization of Somatic Mutations in Cancer Using Network-Based Inference of Protein Activity by Alvarez et. al

• A Single-Cell Tumor Immune Atlas for Precision Oncology by Nieto et. al

• Perspective on Oncogenic Processes at the End of the Beginning of Cancer Genomics by Ding et. al

• Intricate Genetic Programs Controlling Dormancy in Mycobacterium Tuberculosis by Peterson. et al

• Transcriptional Network Modulated by thePprognostic Signature Transcription Factors and Their Long Noncoding RNA Partners in Primary Prostate Cancer by Jiang et. al.

I learned some things from reading these articles:

Firstly, I got an understanding of the way gene regulation works. Only some of the genes in our body actually code for proteins that will be expressed. The other genes are regulatory, meaning that they control whether a different gene will be expressed. Typically, there is a genetic network of such regulatory regions that ultimately control the genes that code for proteins. Transcription factors are the proteins that connect the regulatory regions to the coding regions through various binding sites.

Cancer is the result of errors in the regulation of the cell cycle which result in uncontrolled cell division. There are many, many genes (tumor suppressor genes and proto-oncogenes) that are involved in cancer. Thus, knowing transcription factor binding data can help to predict what exact genes are involved in cancer.

Secondly, I learned that not all transcription factors are created equal. The more binding sites a gene has, the more transcription factors that can bind to it. The more transcription factors that can bind to a cell, the more strongly that gene will be expressed. Aside from binding sites, the activity of a transcription factor impacts gene expression. It doesn’t matter how many binding sites a gene has if the transcription factors are not moving enough to initiate collisions with a binding site. Once transcription factors activity is observed, it can be used to determine which genes are being expressed as a result. Then, the genes can be used to hypothesize patient risk. Basically, the adverse genes that are expressed more are of higher importance for doctors to treat. This can help the researchers to develop targeted treatments, and for clinicians to administer such treatments.

Finally, there is a large precedent for using computer algorithms to map and analyze genome data. Programs like ARACNE can map genes that interact (like genetic switches) and programs like VIPER map transcription factor activity.

But how does this connect to my project?

Since transcription factors can be used to predict what genes have a big impact on cancer, if I can use machine learning to predict what happened transcription factors are most active in breast cancer, I could predict what gene networks are most active. This data can help research design experiments to confirm the findings.

And machine learning is perfect for such a predictive task and I’m planning to analyze the available breast cancer data. I chose breast cancer for a few reasons:

1. There’s a large data set that’s publicly available for breast cancer

2. Breast cancer has a lot of sub-types and not all of them are well understood. While some kinds of breast cancer are tied to known genes (BRCA1 and BRCA2), most breast cancer cases are not identifiable in this way. Gene tests like Oncotype DX and MammaPrint can help doctors decide if a patient needs chemotherapy. But they only apply to some types of breast cancer. And even then, they help in only about 30–40% of those patients.

3. While 60-70% of the breast cancers are detected early, 25-20% of patients with breast cancer see their cancer come back. So, if I find something, it will make a meaningful difference.

My hope is to use a Transcriptional Regulatory Network, a network mapping how transcription factors control genes, to train a machine learning model that can categorize breast cancer patients as high-risk or low-risk for a fast onset or bad cancer outcomes.

By focusing on transcription factors, my goal is that along with predicting risk, my model also predicts the biological reasons behind that risk.

Thank you for reading!

Subscribe now

This blog will chronicle my independent research project over the next 8 months, where I’ll be diving deep into the intersection of genomics, cancer biology, and machine learning. It’s a mix of curiosity, challenge, and the hope that a student like me can contribute — even a little — to the world of science.

What I’m Working On

Over the next year, I’ll be building a machine learning model to try to predict whether a breast cancer patient is high-risk or low-risk for poor outcomes based on the activity of their genes.

Specifically, I’ll be studying something the Transcriptional Regulatory Networks (TRNs): the networks of genes and gene regulators that decide how genes are expressed. In cancer, these networks often break down. My project asks:

“Can we use patterns in these networks to help predict which patients are likely to do worse — and might need more aggressive treatment?”

This idea brings together public cancer data from The Cancer Genome Atlas (TCGA), machine learning tools like Random Forests and XGBoost, and network-based biology using tools like GENIE3 or pySCENIC

I’ll be doing this independently with mentorship, learning the tools of computational biology as I go: from Python and Jupyter notebooks to scientific writing and data visualization.

Why I’m Doing This — and Why Now

I’ve been fascinated by biology since middle school — not just the facts, but the patterns behind them. Over the years, my interests expanded into neuroscience, genetics, and most recently, how complex systems like the human body can be understood with data and models.

But this project is personal, too.

I’ve seen family members affected by cancer. I’ve read the survival statistics. I know that even with decades of progress, we’re still far from being able to predict which patients will respond well and which won’t. And I have kept wondering,

“Could we do better if we looked not just at individual genes or mutations, but at how entire gene networks behave?”

This summer felt like the right time to begin. I’ve just completed AP Biology and some advanced math and coding work. I’m a Junior Health Scholar at a hospital, where I’ve seen firsthand how unpredictable cancer can be. And I want to take all that inspiration and start contributing to science myself.

What This Blog Will Cover

Every week, I’ll be writing a short post:

• What I learned that week

• What I built or coded

• What questions I’m still wondering about

From posts about survival plots or gene networks to complaints about debugging Python code at 11pm, this is where I’ll share the journey.

Future goals

My hopes with this project:

• Learn how real biomedical research works

• Deepen my understanding of cancer biology and AI

• Share this journey through blog posts and visualizations

• Eventually enter a science competition or publish what I find

Whether you’re a student, scientist, mentor, or just curious, I’d love to have you follow along!