Biology subjects

Black, N.

Publications and source records attributed to Black, N..

5 recordsLinked to original sources

bioRxiv2025

Evaluating peer-to-peer bioinformatics education: a case study of student learning outcomes and community impact in an undergraduate multi-omic data analysis course

Computational methodology has become ubiquitous in biomedical research with the rise of big data analysis and popularity of artificial intelligence and machine learning. However, undergraduate bioinformatics education has largely struggled to keep pace with the demand for bioinformatics skills, due to a combination of social and resource-based barriers. In this case study, we discuss the application and outcomes of Multi-Omic Data Analysis, a peer-to-peer learning-based undergraduate bioinformatics course offered by the Department of Quantitative and Computational Biology at the University of Southern California. Over eight semesters, a cohort of student instructors taught 2-3 weekly lectures to 107 undergraduate students in the Quantitative Biology Bachelor of Science degree program. Lectures covered a range of topics, including R and Python data analysis, scientific communication, and general research readiness as undergraduate students. We find that bioinformatics education courses structured around peer-to-peer learning have great potential to overcome many of the obstacles to comprehensive undergraduate bioinformatics education, and provide additional benefits related to student cohesion and community. We further discuss the longevity and feasibility of such courses, both specific to our program and in undergraduate universities at large. Author SummaryThere is a growing need for accessible and beginner-friendly bioinformatics education at the undergraduate level. Traditional coursework often fails to meet this demand due to disciplinary separation between biology and computer science, as well as limited institutional resources. We evaluate a unique approach involving a student-led undergraduate bioinformatics course at the University of Southern California to understand how peer-to-peer learning influences student development in computational biology. We examined over 100 students across eight academic semesters. Students varied in prior coding experience and research exposure. Through survey responses, we found that students left the course feeling more confident and prepared for research in faculty labs. Students reported improved competence in bioinformatics skills compared to before they took the course. We found that a major reason for this success was the peer-to-peer learning format. Students mentioned that learning from fellow students made the material more approachable and created a supportive community where they felt comfortable asking questions. Our work demonstrates that student-led initiatives can be a highly effective solution for filling critical gaps in university curricula, making bioinformatics education more accessible.

scientific communication and education↗

bioRxiv2025

Transfer learning framework via Bayesian group factor analysis incorporating feature-wise dependencies

Transfer learning considers distinct but related tasks defined over heterogeneous domains and aims to improve generalization and performance through knowledge transfer between tasks. This approach can be especially advantageous in biomedical contexts with insufficient labeled training data, where joint learning across domains can enable inference in otherwise underpowered datasets. High-dimensional biomedical data is characterized with redundancy, rendering non-linear dependencies among features. Existing models often fail to leverage such feature dependencies during inference, limiting their ability to model complex biological systems. We present a Bayesian group factor analysis transfer learning framework that supports multitask, multi-modal learning. Our approach learns a shared latent space within each domain, simultaneously across multiple domains, and uses a feature-wise prior to model complex relationships. We evaluate our framework using controlled synthetic data experiments and four disjoint patient cancer datasets from acute myeloid leukemia and neuroblastoma. We show that our method improves drug response prediction and more readily recapitulates consensus biomarkers of drug response. Similarly, our approach improves tumor purity prediction and identifies a robust gene signature associated with it. Our framework is scalable, interpretable, and adaptable across target phenotypes, offering a robust solution for a wide range of heterogeneous multi-omics problems.

bioinformatics↗

bioRxiv2025

Proteomic characterisation of the early rheumatoid arthritis-cardiovascular disease multimorbid axis.

BackgroundInflammation contributes to the increased risk of cardiovascular disease (CVD) observed in people with rheumatoid arthritis (RA), with increased prevalence observed from time of diagnosis. Subclinical vascular and myocardial abnormalities can be detected with cardiovascular magnetic resonance (CMR) imaging in otherwise low risk individuals. Identifying associated circulating markers could enable diagnostics and implicate biological pathways of RA-CVD. The studys objective was to identify blood-based proteins associated with CMR measures of vascular and myocardial abnormality and associated inflammatory and/or cardio-metabolic pathways in a new-onset RA cohort. MethodsSerum samples (baseline/pre-treatment, N = 75 and year 1, N = 71) from CADERA (Coronary Artery Disease Evaluation in Rheumatoid Arthritis) participants, a subgroup of a randomised controlled trial, who underwent CMR, were used to measure 334 proteins across 4 pre-defined Olink panels (Inflammation, Cardiovascular-II, Cardiovascular-III, Cardiometabolic). Bayesian mixed effects regression analyses, unsupervised hierarchical clustering and protein network analyses were applied. Normalised protein expression from 334 Olink proteins were used as exposures in the regression analyses. CMR measures of vascular stiffness (aortic distensibility and stiffness index) and myocardial tissue characteristics (native T1, myocardial extracellular volume and late gadolinium enhancement) were each used as individual outcomes in regression analyses. Associations were considered significant at 95% threshold for credible intervals. An expanded physical protein-protein interaction (PPI) network created using the significant (seed) proteins was subjected to topological and enrichment analysis to identify enriched biological pathways. Results54/334 proteins were associated significantly with CMR measures at baseline (7 - vascular; 48 - myocardial tissue characteristics; Coagulation factor 11 with both). Two proteins, TNFSF13B (B-cell activating factor) and CRTAC1, were associated with CMR measures at baseline, were sensitive to change over time and co-varied with changes in CMR measure. Topological analysis of expanded PPI network revealed GRB2 as most connected (degree score=64; closeness score=118.83) and four key signalling pathways, including JAK-STAT and EGFR tyrosine kinase inhibitor resistance, emerged as significant. ConclusionsThis first proteomic study of treatment-naive early RA and subclinical cardiovascular pathology identifies proteins that could aid diagnostic test development and implicates signalling pathways in the RA-cardiovascular axis to inform on our understanding of the basis of RA-CVD.

systems biology↗

bioRxiv2022

Sex and BDNF Val66Met Polymorphism matter for exercise-induced increase in neurogenesis and cognition in middle-aged mice

Females show greater benefits of exercise on cognition in both humans and rodents, which may be related to brain-derived neurotrophic factor (BDNF). A single nucleotide polymorphism (SNP), the Val66Met polymorphism, within the human BDNF gene, causes impaired activity-dependent secretion of neuronal BDNF and impairments to some forms of memory. We evaluated whether sex and BDNF genotype (Val66Met polymorphism (Met/Met) versus wildtype (Val/Val)) influenced the ability of voluntary running to increase cognition and hippocampal neurogenesis in mice. Middle-aged C57BL/6J (13 months) mice were randomly assigned to either a control or an aerobic training (AT) group (running disk access). Mice were trained on the visual discrimination and reversal paradigm in a touchscreen-based technology to evaluate cognitive flexibility. BDNF Met/Met mice had fewer correct responses compared to BDNF Val/Val mice on both cognitive tasks. Female BDNF Val/Val mice showed greater cognitive flexibility compared to male mice regardless of AT. Despite running less than BDNF Val/Val mice, AT improved performance in both cognitive tasks in BDNF Met/Met mice. AT increased neurogenesis in the ventral hippocampus of BDNF Val/Val mice of both sexes and increased the proportion of mature type 3 doublecortin-expressing cells in the dorsal hippocampus of female mice only. Our results indicate AT improved cognitive performance in BDNF Met/Met mice and increased hippocampal neurogenesis in BDNF Val/Val mice in middle age. Furthermore, middle-aged female mice may benefit more from AT than males in terms of neuroplasticity, an effect that was influenced by the BDNF Val66Met polymorphism. HighlightsO_LIBDNF Met/Met mice performed worse than BDNF Val/Val mice in middle-age C_LIO_LIAerobic training (AT) increased cognitive performance in BDNF Met/Met mice C_LIO_LIAT increased neurogenesis in middle-aged BDNF Val/Val mice only C_LIO_LIFemale BDNF Val/Val mice had better cognitive flexibility than males regardless of AT C_LIO_LIAT increased more mature new neurons in middle-aged female mice C_LI

neuroscience↗

bioRxiv2019

Circadian control of intrinsic heart rate via a sinus node clock and the pacemaker channel

In the human, there is a circadian rhythm in the resting heart rate and it is higher during the day in preparation for physical activity. Conversely, slow heart rhythms (bradyarrhythmias) occur primarily at night. Although the lower heart rate at night is widely assumed to be neural in origin (the result of high vagal tone), the objective of the study was to test whether there is an intrinsic change in heart rate driven by a local circadian clock. In the mouse, there was a circadian rhythm in the heart rate in vivo in the conscious telemetrized animal, but there was also a circadian rhythm in the intrinsic heart rate in denervated preparations: the Langendorff-perfused heart and isolated sinus node. In the sinus node, experiments (qPCR and bioluminescence recordings in mice with a Per1 luciferase reporter) revealed functioning canonical clock genes, e.g. Bmal1 and Per1. We identified a circadian rhythm in the expression of key ion channels, notably the pacemaker channel Hcn4 (mRNA and protein) and the corresponding ionic current (funny current, measured by whole cell patch clamp in isolated sinus node cells). Block of funny current in the isolated sinus node abolished the circadian rhythm in the intrinsic heart rate. Incapacitating the local clock (by cardiac-specific knockout of Bmal1) abolished the normal circadian rhythm of Hcn4, funny current and the intrinsic heart rate. Chromatin immunoprecipitation demonstrated that Hcn4 is a transcriptional target of BMAL1 establishing a pathway by which the local clock can regulate heart rate. In conclusion, there is a circadian rhythm in the intrinsic heart rate as a result of a local circadian clock in the sinus node that drives rhythmic expression of Hcn4. The data reveal a novel regulator of heart rate and mechanistic insight into the occurrence of bradyarrhythmias at night.

physiology↗