Additional file 7: Figure S4. of Classification of low quality cells from single-cell RNA-seq data
Year: 2016
Type: article
Abstract: Datasets distant from mES training data. (A) Comparing log normalized UMI counts (y-axis) and log normalized read counts (x-axis) for each gene in 960 mESCs. (B) PCA of first two principal components of all features. Low quality cells separate from high quality cells. (C, D) PCA plot of features of two published human cancer cell datasets [28, 53]. Boxplots on the left and bottom show the top three features separating low from high quality cells for PC1 and PC2, respectively. They align with our previous findings that the mtDNA and ERCC to mapped reads ratios are upregulated in low quality cells. (E) Feature-based PCA combining mouse ES training set and two published human cancer datasets. â Cytoplasmâ separates not only the human from the mouse but also the two different cancer samples from each other, meaning that the features trained on mouse cells are not directly transferrable to human cancer cells. (PDF 591Â kb)
Source: Figshare
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7136280348