General
Recent advances in single-cell and spatial -OMICS technologies are transforming the life sciences, resulting in increasingly high-dimensional datasets. While highly informative, such data are challenging to interpret directly, creating a strong demand for dimensionality reduction techniques.
This Adore seminar aims to provide a foundational and critical understanding of dimensionality reduction methods in life-sciences research. The seminar consists of three parts:
1. An introduction to high-dimensional data and the basics of linear dimensionality reduction
2. The algorithmic foundations of widely used non-linear techniques (tSNE and UMAP), explained for entry-level life scientists
3. A critical discussion of advantages and limitations of non-linear dimensionality reduction, including scalability, faithful data representation, interpretability, and common pitfalls, complemented by practical recommendations and examples of misuse in life-sciences research.