bioRxiv Science⌕ Search

Biology subjects

Banjac, J.

Publications and source records attributed to Banjac, J..

2 recordsLinked to original sources

Deep learning for classifying neuronal morphologies: combining topological data analysis and graph neural networks

Neuronal shape determines how neurons process and integrate information, yet a consistent and objective classification of neuronal morphologies remains elusive. Current approaches rely heavily on subjective expert views or on predefined features, limiting reproducibility and interpretability. Here, we present an interpretable deep learning framework that unifies topological data analysis, graph neural networks, and traditional morphometrics to classify neuronal morphologies objectively and transparently. Our framework compares complementary mathematical representations of neurons to capture geometric, topological, and graph-structural information. Then it benchmarks their performance against expert-labeled datasets. We show that topology- and graph-based models achieve accuracies comparable to human experts, revealing that both global branching invariants and local connectivity patterns are essential to define morphological cell types. Using explainable artificial intelligence methods, we identify structural features driving each classification decision, bridging computational and neuroanatomical interpretations. This open source and reproducible approach provides a foundation for scalable, interpretable and biologically meaningful neuronal taxonomy, enabling consistent comparisons between data sets and species.

neuroscience↗

Microbiome Toolbox: Methodological approaches to derive and visualize microbiome trajectories

SummaryThe gut microbiome changes rapidly under the influence of different factors such as age, dietary changes or medications to name just a few. To analyze and understand such changes we present a microbiome analysis toolbox. We implemented several methods for analysis and exploration to provide interactive visualizations for easy comprehension and reporting of longitudinal microbiome data. Based on abundance of microbiome features such as taxa as well as functional capacity modules, and with the corresponding metadata per sample, the toolbox includes methods for 1) data analysis and exploration, 2) data preparation including dataset-specific preprocessing and transformation, 3) best feature selection for log-ratio denominators, 4) two-group analysis, 5) microbiome trajectory prediction with feature importance over time, 6) spline and linear regression statistical analysis for testing universality across different groups and differentiation of two trajectories, 7) longitudinal anomaly detection on the microbiome trajectory, and 8) simulated intervention to return anomaly back to a reference trajectory. Availability and implementationThe software tools are open source and implemented in Python. The link to the interactive dashboard is https://microbiome-toolbox.herokuapp.com/. For developers interested in additional functionality of the toolbox, the Python package can be downloaded from https://pypi.org/project/microbiome-toolbox/. The toolbox is modular allowing for further extension with custom methods and analysis. The code is available on Github https://github.com/JelenaBanjac/microbiome-toolbox. ContactShaillayKumar.Dogra@rd.nestle.com Supplementary InformationSupplementary data are available at Bioinformatics online.

bioinformatics↗