Toggle Accessibility Tools

Lotreck, Serena

Plant Biology / CMSEhome icon
lotrecks@msu.eduenvelope icon
Lab: Shin-Han Shiu
Research interests: Using machine learning interpretability on model systems to study genetic interactions; information extraction and knowledge graph construction from plant biology literature.

Quick Profile

What got you interested in plants and plant science?
During my freshman year in college, I had to take a comparative physiology class, and one half of the class was taught by a plant physiologist. His PowerPoint slides were just images of plants and diagrams of physiological processes, and he would talk so enthusiastically about all the material that I couldn't help but be interested. When it came time to look for research labs, I decided I would look for plant biology labs to see what it was all about, and ended up joining a lab that studied plant-insect interactions. The rest was history!

What is your research about?
I study the inner workings of machine learning models that make accurate predictions about genetic phenomena. By studying how models make decisions, we can untangle complex molecular and genetic relationships that give rise to the phenomena we predict. This involves building models to predict answers we already know, so that we can then get under the hood of the model and ask, how did we make this prediction? 

Additionally, I work on creating comprehensive knowledge maps of the plant science literature, in order to facilitate scientists' literature search efforts by placing more comprehensive and digestible material within their reach, as well as to explore the existence of contradictory facts within published literature. 

What is the potential societal impact of your research?
Understanding genetic relationships can have many applications; one example is that understandings about plant genetics can be directly applied in breeding settings to improve crop varieties.

Creating search tools to aid scientists in getting a big-picture view of all research that has been done in a given field will help scientists generate better hypotheses and avoid redundantly investigating things that have already been studied. Additionally, by studying the meta-structure of these graphs, we can address why and how conflicting information appears in the literature, with the end goal of helping the scientific literature body provide more accurate and consistent information. 

Where do you see yourself in 10 years?
In ten years, I hope to be working as a data scientist in a start-up or small company that works on crop improvement, sustainable agriculture, or algal biofuels. 

On a Saturday afternoon, you'll likely find me:
Climbing, camping, hiking, or otherwise being eaten alive by mosquitos in pursuit of the outdoors.