Natalie Stanley

MICRON identifies outcome-linked immune microenvironments from spatial imaging proteomics data, improving prognostic and diagnostic prediction in brain cancer.

This paper uses optimal transport to map CD8 T cell differentiation across time and tissues, revealing distinct migration waves and gene regulatory programs.

The paper identifies immune coordination patterns in ALS patients and develops machine learning models to predict disease onset and progression using mass cytometry data.

The paper uses single-cell transcriptomics to study stress hormone effects on immune cells and identifies cell-type-specific changes linked to stress-related diseases.

The paper develops microglia aging clocks using scRNA-seq data and shows frequency-based featurization balances accuracy, interpretability, and efficiency.

Cellograph uses graph neural networks to analyze multi-condition single-cell RNA-seq data, quantifying perturbation effects at single-cell resolution and outperforming existing methods.

This study shows TDP-43 pathology activates immune cells in ALS, linking it to immune dysfunction and suggesting immune monitoring as a diagnostic tool.

This paper benchmarks differential abundance methods to identify condition-specific cell subpopulations in multi-sample single-cell datasets, finding dataset-specific recommendations for method usage.

scLKME generates multi-cellular sample embeddings from single-cell data by using a landmark-based kernel mean embedding method, outperforming existing techniques in robustness and accuracy.

The paper investigates how sketching and graph generation methods affect the identification of biologically meaningful cell-populations in single-cell data, finding that sketching is more critical than graph construction.

This paper introduces DELVE, a feature selection method that improves trajectory inference by identifying dynamically expressed features that better represent biological processes in single-cell data.

This paper introduces cytocoarsening, a graph-coarsening algorithm that compresses single-cell data representations while preserving biological insights from phenotypic and experimental attributes.

This study benchmarks integration methods for temporal single-cell gene expression data to improve trajectory inference and disease prediction.

CytoSet predicts clinical outcomes from cytometry data using set-based deep learning, achieving state-of-the-art performance and robustness across patient samples.

FastPG improves single-cell clustering speed by 27x while maintaining accuracy over existing methods like FlowSOM and PARC.

The paper introduces a machine learning method that integrates immunological knowledge to improve predictive power from mass cytometry data.