Dimensionality Reduction in the Cytobank Platform
In the Cytobank platform, the dimensionality reduction suite is a powerful way for exploratory data analysis and data visualization. The suite now contains four dimensionality reduction algorithms that can reduce high-dimensional data to two dimensions for easy visualization:
- viSNE/tSNE1 (a non-linear dimensionality reduction algorithm developed based on Stochastic Neighbor Embedding)
- tSNE-CUDA2 (a state-of-the-art implementation of the t-SNE algorithm)
- UMAP3 (a non-linear dimension reduction algorithm)
- opt-SNE4 (a t-SNE based algorithm that can automatically optimize the early exaggeration process and the learning rate value for a t-SNE analysis run)
In cytometry data analysis, researchers usually run these dimensionality reduction algorithms after compensating, scaling and pre-gating the data.
Although each of the algorithms functions in a similar way — i.e., transforming data from a high-dimensional space into a low-dimensional space so that the low-dimensional representation retains some meaningful properties of the original data — they do differ slightly in how they go about reducing dimensionality.
How to Compare Results of Different Dimensionality Reduction Algorithms
In this video, Jason Emo — an application scientist here at Beckman Coulter Life Sciences — shows you how to use the tools in the new dimensionality reduction suite to analyze your high-dimensional cytometry data, as well as how to compare results of different dimensionality reduction algorithms in the Cytobank platform.
Cytobank Free TrialCytobank is a cloud-based platform for the analysis, storage, and sharing of flow and mass cytometry data. It offers machine learning-assisted analysis of high-dimensional, single-cell data and is designed to let you easily collaborate with colleagues from different departments and regions from any web-enabled device.
- L.J.P. van der Maaten, Journal of Machine Learning Research, 2014
- Chan DM et al., Journal of Parallel and Distributed Computing, 2019
- orey J. Nolet. et al., arXiv, 2020, arXiv:2008.00325 http://arxiv.org/abs/2008.00325
- Belkina A. et al., Nat Communications, 2019