Inspired by my final project topic question for the Spring 2017 semester on Game Theory, I started working on a personal research on biological networks. I studied how cooperation and defection evolves in a biological network. Biological networks are practically any network involving biological organisms interacting in some way. Evolutionary theorists have long pondered mechanisms for increasing cooperation in human societies (a type of biological network), but although a lot of research has been done on biological networks, it is still a grey area. I designed an agent-based model using Python to simulate a biological network and interactions based on imitate-the-best strategies to understand how cooperation or defection could evolve in such networks. Agents in my model played 2 X 2 Prisoner's Dilemma games with other agents in their network. I was able to discover very interesting, complex associations. My paper is published in Scientific Reports, a Nature journal, and is available for read here. You can also find it on Pubmed.

Estimating causal inferences has been a constraint to statisticians because of the problem of counterfactuals. The best way to understand what a counterfactual is to consider the outcome of a particular action, at any particular time and condition, and comparing that against an entirely different action under the same conditions and time. Now this is practically impossible because there is no way of reenacting exactly the same conditions as those for the initial action. To get around this problem, the matching technique was introduced. With matching, one creates a control and treatment group. A member in the control group would be paired with another member from the treatment group, who would be very similar to the member from the control group in almost all respects aside from the action. These paired groups are compared in estimating causal inferences of several actions. The problem with this approach is that there might not be very similar pairings, and this could be a huge constraint to approximating the actual causal inference. My project on this was to think about a different approach of estimating causal inferences, while minimizing the error and maximizing accuracy. My solution around this was to train a good regression model that can very well predict out-of-sample data, and given that out counterfactuals fall within the convex hull determined by The Gower's distance, then we can make extrapolations to predict the outcome of our counterfactuals. This is circumvents the constraint of finding a good pair for matching a particular datapoint, also, if you have a good regression model, then you can very closely predict the true counterfactual. My project on this was done with R and is available on my GitHub page.

In the summer of 2017, I interned at Pennsylvania State University College of Medicine, Hershey, Pennsylvania, USA as a Summer Undergraduate Researcher. This is a top tier research institute and I was privileged to work in one of the most interesting labs at Penn State. My research was a follow-up on previous research carried out my lab. Previous research has been done on the role of REDD1 (Regulated in Development and DNA Damage response 1) expression in causing retinopathy in diabetic patients. Our results suggested that REDD1 causes blindness in diabetic patients due to diabetes-induced hyperglycemia. My research focused on the effect of REDD1 on liver autophagy. Practically, I measured autophagy markers, majorly ratios of LC3 A/B, in REDD1 wildtype hepatocytes as compared to that in CRISPR REDD1 knockout liver cells after some hours of stimulation with treatment, be it serum deprivation or tunicamycin. With the recent projects I am working on moringa oleifera, especially as it concerns my startup, Miracle Tea Co., I am beginning to look into how moringa oleifera affects the expression of REDD1 and LC3 markers.

Alongside another colleague, I was invited by the Ugandan Mathematical Society (UMS) to Uganda to train participants competing at the International Mathematical Olympiad (IMO). The International Mathematical Olympiad is an international competition where countries choose six exceptionally talented high school students to represent the country at an international mathematics competition. These students are taught advanced concepts in mathematics, such as Combinatorics, Algebra, Geometry, Number Theory and Inequalities. We trained the Ugandan team for the IMO at Makarere University, Kampala, Uganda during the summer of 2016. The team went on to make their first Honorary Mention for the country at the International Mathematical Olympiad. Here is one of the stories that appeared on the Ugandan newspaper. I have since been a trainer for the IMO Ugandan team. I went on ahead that summer to train the contestants from Nigeria at the National Mathematical Center (NMC), Abuja, Nigeria who were also going to compete that year at the IMO. They were also able to secure an Honorable Mention at the competition. Learn more about the IMO. Visit the official website of the International Mathematical Olympiad (IMO).

I train Nigerian participants who compete at the International Biology Olympiad (IBO). The IBO is a prestigious international competition where each country is represented by their best four high school students to compete in biology. As a former two-time representative for Nigeria and gold medalist in the International Science Olympiad (ISO), I was invited to train the team to compete at the IBO for years 2014 and 2016. See Resume for more information. The arrangement for the training was organized by National Mathematica Center (NMC), Kwali, Abuja, Nigeria. Learn more about the IBO. Visit the official website of the International Biology Olympiad (IBO).