inst/www/helpfiles/choose_cluster.md

In ChromSCape: Analysis of single-cell epigenomics datasets with a Shiny App

PCA - Principal Component Analysis

This linear dimensionality reduction method is applied to the normalized count matrix. In this plot, each cell is a dot. The cells are placed according to their projection in the Component_1 & Component_2 (you are able to select any components). Cells close together have similar epigenomic traits. Coloring by 'total_counts' is a good way to ensure there is no direct correlation between the library size and the first components. Among other things, applying PCA to the large count matrix allows to:

• Overcome the Curse of Dimensionality for downstream clustering of cells.
• Eliminate redudancy between multiple features & reduce noise
• Vizualise high-dimensional dataset in lower 2D space

ChromSCape documentation built on Nov. 8, 2020, 6:57 p.m.