Streptococcus mutans strain 3668 causes caries, a form of tooth decay known as cavities. To prevent caries, phages (viruses that infect and kill bacteria) can be used. Unlike antibiotics, phages can kill specific strains of bacteria while leaving helpful microbiota unaffected. To identify phages in nature that may infect S. mutans, we used two experimental approaches: (1) spot assay, and (2) plaque assay. Both methods take advantage of being able to easily visualize a “plaque”, which is defined as being an area of bacterial death caused by infection of a bacterial cell by a single phage and subsequent cell lysis. In this study, we tested two host bacteria (Escherichia coli and S. mutans), and their abilities to be infected by either known phage (MS2, infects E. coli: positive control), or environmental samples. We successfully grew plaques for MS2 (E. coli host), and for one set of environmental sampling on aquariums found a single sample that caused an area of clearance on S. mutans bacterial lawns. Follow-up experiments suggest that the environmental sample that killed S. mutans was either an antibiotic (not a phage), or that the initially infecting phage became deactivated within 24 hrs, preventing further propagation and purification. Future studies from alternative environmental samples, utilizing both the spot and plaque assays, will be conducted to find novel phages that can be used to treat caries.

The process of aging has long been an active topic of research; however, methods of studying the aging process have not yielded conclusive results. In our work, we plan to utilize drugs and their effects on C. elegans to study the aging process. In particular, we focus on compounds that extend C. elegans lifespan (called NLEs, or Nematode Life Extending Compounds). However, due to the wide scope of compounds that exist, it would be inefficient and prohibitively expensive to screen them all in vivo. There has currently been no research into the efficacy of different machine learning models to screen for life-extending compounds in C. elegans. Therefore, we plan to utilize machine learning models to predict these NLEs, effectively narrowing the chemical search space for cheaper and more efficient screens. We began by mining around 130 NLEs from various literature sources to serve as positive samples, then combined them with a dataset of assumed negative samples. The Morgan Fingerprints of these compounds were used as features to train our models. We will implement a variety of machine learning techniques to narrow down the scope of our compounds, and the robustness of our models will be evaluated and compared to each other using ROC and precision-recall curves. Some models we will be testing are support vector machines (SVMs), gradient boosting models via XGBoost, and MLP neural networks. These methods will be evaluated using 10-fold cross-validation. Results will be available Summer 2024.

Since the discovery of cancer stem cells in 1994 and the discovery of embryonic stem cells in 1998, research on these subjects has been rapidly expanding (Lapidot et al., 1994; Thompson et al., 1998). With the newer tool that is the epigenetic field, significant stem cell research currently focuses on the influence of epigenetic markings, such as acetylation— the addition of an acetyl group, “a small molecule made of two carbon, three hydrogen, and one oxygen atoms,” to DNA (National Cancer Institute). The proposed research deals with histone proteins, proteins around which DNA is wrapped, and the specific amino acid lysine. The purpose is to further understand the consequences of acetylating the 56th lysine on histone protein H3 regarding the cells’ ability to differentiate and to multiply indefinitely. Mutant (cells with emulated acetylated or unacetylated H3K56) and wild type cells will be analyzed utilizing mCherry-tagging, Benzonase Nuclease digestion, SDS-PAGE, and LS-MS coupled with online databases. The analysis will focus on protein composition and histone modifications of each type of cell. The expected outcome will be for unacetylated mutants to have a lower concentration of proteins that are related to stemness, showing that the acetylation of H3K56 may play a fundamental role in supporting stem cell qualities (Swain et al., 2020). The proposed research will be helpful for further experiments done on stem cells, embryonic, cancer, or induced, and could potentially lead to breakthroughs in cancer therapy by either proving or disproving a connection between H3K56ac and cell stemness.

Polycystic ovary syndrome (PCOS) is a common reproductive endocrine disorder characterized by high testosterone (T), menstrual irregularities, polycystic ovaries and cardiometabolic disease. Heritable factors significantly influence PCOS development, but contributing gene variants are not well understood. This study aimed to identify heritable traits contributing to PCOS using data from a cohort of female rhesus macaques at the Wisconsin National Primate Research Center with naturally high T or defined PCOS phenotypes. 34 rhesus macaques were chosen for this study, 14 being identified as having high testosterone (greater than or equal to 0.31 ng/mL), and 20 being identified as having normal testosterone (less than 0.31 ng/mL). Average menstrual cycle intervals were calculated, blood samples were obtained to measure varying hormone values and ovarian follicles were quantified to identify polycystic ovary morphology. Of five significantly heritable traits, two related to T were over-represented among PCOS monkeys, T to androstenedione (A4) ratio and A4 to DHEA ratio. We therefore hypothesize that female macaques exhibiting PCOS will carry high combined annotation dependent depletion (CADD) scores for gene variants in PCOS risk genes compared to macaques without PCOS. This research aims to provide insights into genetic factors contributing to PCOS and potential therapeutic targets or biomarkers.