Explore the vast range of titles available.

Understanding which microbes differ between groups of interest could reveal therapeutic targets and diagnostic biomarkers. However, current analysis methods treat all microbes equally (similar to using the same ruler to measure everything, regardless of what matters most). This means subtle but biologically important differences may go undetected, especially when only a few key species drive disease states while hundreds of “bystander” species add noise. Metric Learning for Statistical Inference (MeLSI) solves this by learning which microbes matter most for each specific comparison. In comparing male and female gut microbiomes, MeLSI identified specific bacterial families driving the differences, providing actionable biological insights that standard methods miss. This capability is particularly crucial for detecting early disease biomarkers, where differences are subtle and masked by biological variability. By telling researchers not just whether groups differ, but which specific microbes drive those differences, MeLSI accelerates the path from microbiome data to testable biological hypotheses and clinical applications.

Background: Alzheimer’s disease (AD) is a leading cause of dementia, and it is significantly influenced by the apolipoprotein E4 (APOE4) gene and gender. This study aimed to use machine learning (ML) algorithms to predict brain age and assess AD risk by considering the effects of the APOE4 genotype and gender. Methods: We collected brain volumetric MRI data and medical records from 1100 cognitively unimpaired individuals and 602 patients with AD. We applied three ML regression models—XGBoost, random forest (RF), and linear regression (LR)—to predict brain age. Additionally, we introduced two novel metrics, brain age difference (BAD) and integrated difference (ID), to evaluate the models’ performances and analyze the influences of the APOE4 genotype and gender on brain aging. Results: Patients with AD displayed significantly older brain ages compared to their chronological ages, with BADs ranging from 6.5 to 10 years. The RF model outperformed both XGBoost and LR in terms of accuracy, delivering higher ID values and more precise predictions. Comparing the APOE4 carriers with noncarriers, the models showed enhanced ID values and consistent brain age predictions, improving the overall performance. Gender-specific analyses indicated slight enhancements, with the models performing equally well for both genders. Conclusions: This study demonstrates that robust ML models for brain age prediction can play a crucial role in the early detection of AD risk through MRI brain structural imaging. The significant impact of the APOE4 genotype on brain aging and AD risk is also emphasized. These findings highlight the potential of ML models in assessing AD risk and suggest that utilizing AI for AD identification could enable earlier preventative interventions.

Background Apolipoprotein ε4 allele (APOE4) is the strongest genetic risk factor for late‐onset Alzheimer’s disease (AD). Compared with non‐carriers, cognitively normal APOE4 individuals have shown brain atrophy and lower cerebral blood flow (CBF) decades before AD pathological and clinical symptoms appear. The goal of the study is to determine if using Sirolimus, an FDA‐approved mTOR inhibitor, could restore the brain volumes in structures related to cognitive functions and global CBF (gCBF) for asymptomatic APOE4 carriers compared with non‐carriers. Method The study was performed at the University of Missouri. Low dose Sirolimus (1 mg/day) was given for 4 weeks to three middle‐aged, cognitively normal APOE4 carriers (F:M = 2:1) and two female APOE3 individuals (45‐65 yrs; MOCA > 27). Oral swabs were used to determine APOE status. 3T MRI‐based structural MPRAGE and T1‐weighted images, and pseudo‐continuous arterial spin‐labeled (PCASL, 1.7x1.7x4mm resolution) were acquired at baseline (pre‐treatment) and Post‐Sirolimus (the end of treatment). FreeSurfer 7.4 was used to automatically generate segments for volumetric analysis and PCASL images were analyzed via MANGO software for perfusion weighted gCBF. Result APOE4 carriers had significantly increased Caudate and Hippocampal volumes (Figure 1A) and gCBF (Figure 1B) after 4 weeks of Sirolimus treatment. These differences were not found in the non‐carriers (APOE3 participants). The quantitative data is shown in Table 1. It also shows that APOE4 carriers had significantly lower Hippocampal volume and gCBF at baseline compared with that of APOE3 participants (indicated by “**”), and Sirolimus tends to restore the values to closer to those of the non‐carriers. No side effects were observed, and no changes in blood glucose and HbAc1 levels were found in all the participants. Conclusion We show that short‐term Sirolimus treatment can effectively restore brain volumes and gCBF for asymptomatic APOE4 carriers. Specifically in the caudate and hippocampus, which play many roles integral in normal cognitive functioning including memory, learning, and emotional regulation, has considerable atrophy in AD. The findings imply that Sirolimus may be useful and effective to mitigate or prevent AD developments for asymptomatic APOE4 carriers.

Background Apolipoprotein e4 allele (APOE4) is the strongest genetic risk factor for late‐onset Alzheimer’s disease (AD). Compared with non‐carriers, cognitively normal APOE4 individuals have shown brain atrophy and lower cerebral blood flow (CBF) decades before AD pathological and clinical symptoms appear. The goal of the study is to determine if using Sirolimus, an FDA‐approved mTOR inhibitor, could restore the brain volumes in structures related to cognitive functions and global CBF (gCBF) for asymptomatic APOE4 carriers compared with non‐carriers. Method The study was performed at the University of Missouri. Low dose Sirolimus (1 mg/day) was given for 4 weeks to three middle‐aged, cognitively normal APOE4 carriers (F:M = 2:1) and two female APOE3 individuals (45‐65 yrs; MOCA > 27). Oral swabs were used to determine APOE status. 3T MRI‐based structural MPRAGE and T1‐weighted images, and pseudo‐continuous arterial spin‐labeled (PCASL, 1.7×1.7×4mm resolution) were acquired at baseline (pre‐treatment) and Post‐Sirolimus (the end of treatment). FreeSurfer 7.4 was used to automatically generate segments for volumetric analysis and PCASL images were analyzed via MANGO software for perfusion weighted gCBF. Result APOE4 carriers had significantly increased Caudate and Hippocampal volumes (Fig. 1A) and gCBF (Fig. 1B) after 4 weeks of Sirolimus treatment. These differences were not found in the non‐carriers (APOE3 participants). The quantitative data is shown in Table 1. It also shows that APOE4 carriers had significantly lower Hippocampal volume and gCBF at baseline compared with that of APOE3 participants (indicated by “**”), and Sirolimus tends to restore the values to closer to those of the non‐carriers. No side effects were observed, and no changes in blood glucose and HbAc1 levels were found in all the participants. Conclusion We show that short‐term Sirolimus treatment can effectively restore brain volumes and gCBF for asymptomatic APOE4 carriers. Specifically in the caudate and hippocampus, which play many roles integral in normal cognitive functioning including memory, learning, and emotional regulation, has considerable atrophy in AD. The findings imply that Sirolimus may be useful and effective to mitigate or prevent AD developments for asymptomatic APOE4 carriers.

Microbiome beta diversity analysis relies on distance-based methods including PERMANOVA combined with fixed ecological distance metrics (Bray-Curtis, Euclidean, Jaccard, and UniFrac), which treat all microbial taxa uniformly regardless of their biological relevance to community differences. This \"one-size-fits-all\" approach may miss subtle but biologically meaningful patterns in complex microbiome data. We present MeLSI (Metric Learning for Statistical Inference), a novel machine learning framework that learns data-adaptive distance metrics optimized for detecting community composition differences in multivariate microbiome analyses. MeLSI employs an ensemble of weak learners using bootstrap sampling, feature subsampling, and gradient-based optimization to learn optimal feature weights, combined with rigorous permutation testing for statistical inference. The learned metrics can be used with PERMANOVA for hypothesis testing and with Principal Coordinates Analysis (PCoA) for ordination visualization. Comprehensive validation on synthetic benchmarks and real datasets shows that MeLSI maintains proper Type I error control while delivering competitive or superior F-statistics when signal structure aligns with CLR-based weighting and, crucially, supplies interpretable feature-weight profiles that clarify which taxa drive group separation. On the Atlas1006 dataset, MeLSI achieved stronger effect sizes than the best traditional methods, and even when performance was comparable, the learned feature weights provided biological insight that fixed metrics cannot supply. MeLSI therefore offers a statistically rigorous tool that augments beta diversity analysis with transparent, data-driven interpretability. Understanding which microbes differ between groups of interest could reveal therapeutic targets and diagnostic biomarkers. However, current analysis methods treat all microbes equally (similar to using the same ruler to measure everything, regardless of what matters most). This means subtle but clinically important differences may go undetected, especially when only a few key species drive disease while hundreds of \"bystander\" species add noise. MeLSI solves this by learning which microbes matter most for each specific comparison. In comparing male and female gut microbiomes, MeLSI identified specific bacterial families driving the differences, providing actionable biological insights that standard methods miss. This capability is particularly crucial for detecting early disease biomarkers, where differences are subtle and masked by biological variability. By telling researchers not just whether groups differ, but which specific microbes drive those differences, MeLSI accelerates the path from microbiome data to testable biological hypotheses and clinical applications.

Cell surfaces feature abundant glycosylation. Glycans adorn roughly 90% of cell surface proteins and a huge array of phospholipids and sphingolipids comprising the plasma membrane, and myriad polysaccharides bind to cell surface proteins. This collection of glycans concentrated on cell surfaces is known as the glycocalyx and every cell has one. Like many cellular phenomenon, malignant transformation often results in reliable patterns of changes to the glycocalyx. Just as oncogenes are often upregulated in cancers and tumor suppressor genes downregulated, large cell-surface glycoproteins are very often upregulated as well. And tumor cells are often found to be hypersialylated. So consistently upregulated are the mucin class of glycoproteins in cancer, that they are used as biomarkers of the disease. Clearly, investigation into the role of glycans in oncogenesis is warranted. For decades, these changes were documented correlatively, but a potential causative role in oncogenesis for these glycans remained elusive. The reasons for our lack of understanding are technical in nature; glycans and glycoproteins are very challenging to study using standard molecular biology techniques. Cancer biologists and cell biologists have been able to use molecular genetics techniques to tease apart the structure-function relationship of most proteins they set their sights on, but glycans are not template encoded. While a protein can be manipulated by changing its corresponding genetic DNA, glycans have no such coding counterpart. Glycosylation patterns are the result of countless metabolic pathways and the balance of activity of glycosyltransferases and glycosidases. So there is no such analogous facile manipulation-to-phenotype technique yet available to glycobiologists. Chemical biology has emerged as the savior to the struggling glycobiologist. By utilizing chemical techniques, glycobiologists now have techniques for gaining traction on some cell-surface glycosylation patterns. Our lab developed just such a technique in the construction of our mucin-mimetic glycopolymers. Through chemical synthesis, we can create glycan-containing molecules that emulate the structure of native glycoproteins. Through a hydrophobic tail, these molecules spontaneously insert into the membranes of cells and thus decorate them with the glycans we chose to incorporate into the molecule. In this way we can begin to ask and answer questions such as: what effect does glycan ‘X’ have upon cancer progression when overexpressed on tumor cell surfaces? In chapter 1, I describe my contribution to these chemical tools. While profoundly useful, the early generation of these mucin-mimetic glycopolymers are constitutively uptaken by the cells which they decorate, meaning there is a limited timespan for the types of experiments they can contribute to. I constructed series of various lipid anchors for the glycopolymers and tested their ability to display the glycopolymers on cell surfaces. Using this empirical approach, we discovered that the synthetic sterol cholesterylamine is capable of promoting the recycling of our glycopolymers after endocytosis and thus display them for days to weeks. We go on to show that these mimics of mucin structure are capable of endowing cancer cells with metastatic-like phenotypes in a zebrafish model of metastasis. In chapter 2 we put these molecules to use and answer the question: what contribution to metastasis do mucin glycoproteins make on tumor cell surfaces? Mucins are overexpressed on a massive number of cancers. It is estimated that 65% of all tumors diagnosed each year in the US overexpress MUC1—just one member of this family of cell-surface glycoproteins, of which there are at least twenty! They are found to positively correlate with metastasis, but the mechanism for such a contribution has remained entirely elusive. We found—using our mucin-mimetic glycopolymers and confirmed with genetic MUC1 constructs—that the mucin ectodomain has a unique contribution to the adhesion of cancer cells in the metastatic niche. When a cell arrives at a distal site in order to form a metastatic lesion, it must not only survive there but also proliferate. The cell may very well find that there isn’t much adhesive support for such behavior. In our models, overexpressing mucin ectodomains allows it to grow and survive despite that shortcoming. This work not only provides the first mechanistic support of a role for the mucin ectodomain in metastatic spread, but it has dramatic consequences for the targeting of mucin oncogenes. Work has been done to target the biochemically active cytoplasmic tails of mucins with novel therapeutics. Our studies imply that such targeting is futile and may not combat the mucin’s most deadly structural contribution to oncogenesis: its ectodomain. So how, then, are biomedical scientists to produce drugs which might combat this dramatic contribution of the cancer cell glycocalyx to progression of the disease? In chapter 3, I describe a radically new approach to the treatment of cancer—but also to medical therapies in general. Using a handful of chemical biology techniques, we synthesized a conjugate of a monoclonal antibody and an enzyme. The enzyme can exert its effect on the cell surface once brought into proximity by the antibody. In this way, enzymes that modify the glycocalyx in an anti-cancer manner can be directly targeted to tumors. As a proof of principle, we first made sialidase-trastuzumab conjugates. High expression of sialic acids has long been associated with cancer, but it was only recently that their role in immune-evasion has been appreciated. By displaying a lot of sialic acid on their cell surfaces, tumors are able to bind to inhibitory receptors on immune cells—which might otherwise kill the cancer cells—and avoid their destructive powers. Upon removing those sugars, we have seen that they become susceptible to the immune cells once again. Therefore, in this chapter I demonstrate how an enzyme that cleaves sialic acids can be selectively targeted to tumor cells bearing the antigen of the antibody, and their killing by immune cells enhanced as much as 2 fold, via this new class of therapeutics: the antibody-enzyme conjugates.

The purpose of this causal comparative study is to test the theory of no significant difference that compares pre- and post-test assessment scores, controlling for the instructional delivery model of online and face-to-face students at a Mid-Atlantic university. Online education and virtual distance learning programs have increased in popularity and enrollment since their inception. Students tend to enroll in online courses for their flexibility and convenience and find online courses to be just as challenging as face-to-face courses (Pastore & Carr-Chellman, 2009). Russell (1999) conducted a meta-analysis which found that there were no significant differences between the modes of class delivery on student achievement and learning. Current research supports this analysis; it has been shown that students and instructors perceive online learning to be just as effective as face-to-face (Liaw, Huang, & Chen, 2007). Bloom's Taxonomy has been used to structure the thinking process in education. Elevating an awareness of pedagogical shifting across delivery models will likely lead to more effective university teaching in both face-to-face and distance programs (Girod & Wojcikiewicz, 2009). Utilizing an ANCOVA, research was conducted pre and post instruction that determined differences existed based on the instructional delivery model in a remedial English course favoring face-to-face instruction. Further, regarding the occurrence of higher order thinking skills, statistical analysis based on a t-test indicated that online students more frequently exhibit this skill versus students enrolled in face-to-face instruction.