UMAP Basics for Cytometry скачать в хорошем качестве

UMAP Basics for Cytometry

Understanding UMAP for High-Dimensional Cytometry in Drug Development This webinar covers the basics of Uniform Manifold Approximation and Projection (UMAP) for interpreting high-dimensional cytometry data. It compares UMAP with Principal Component Analysis (PCA) and demonstrates how UMAP can more effectively distinguish between distinct cell populations in 2D visualizations. Using examples from published research and real datasets, the tutorial explains the clustering and dimensionality reduction processes, highlighting the advantages of UMAP for identifying rare or subtle cell populations relevant to drug response and toxicity. Ideal for drug developers looking to visualize complex data sets. 00:00 Introduction 00:39 Reminder about Principal Component Analysis (PCA) 01:19 Comparing PCA and UMAP Visualizations 03:56 Deep Dive into UMAP 05:40 Real-World Applications of UMAP 06:40 Simplifying UMAP with Fashion MNIST 08:24 Applying UMAP to Cytometry Data 09:10 Conclusion and Benefits of UMAP References: Dimensionality reduction for visualizing single-cell data using UMAP, Newell et al: https://www.nature.com/articles/nbt.4314 UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction, McInnes et al: https://arxiv.org/abs/1802.03426