id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
10_1101-2021_02_09_430550	Song, Dongyuan	scPNMF: sparse gene encoding of single cells to facilitate gene selection for targeted gene profiling	2021	37	.pdf	application/pdf	13512	1548	64	(scPNMF) method to select informative genes from scRNA-seq data in an unsupervised way. Therefore, for scRNA-seq data analysis, informative gene selection Besides scRNA-seq data analysis, informative gene selection is also crucial for designing number and a scRNA-seq dataset, scPNMF selects informative genes based on its weight matrix; First, the informative genes selected by scPNMF lead to the most accurate cell clustering. the informative genes and weight matrix of scPNMF lead to the best cell type prediction accuracy Figure 3: Benchmarking scPNMF against 11 informative gene selection methods on seven scRNA-seq (b) UMAP visualization of cells in the Zheng4 dataset based on 100 informative genes selected by We benchmark scPNMF against the 11 gene selection methods in terms of cell type prediction We propose scPNMF, an unsupervised gene selection and data projection method for scRNA-seq For cell type prediction, we project every targeted gene profiling dataset and its scRNA-seq	./cache/10_1101-2021_02_09_430550.pdf	./txt/10_1101-2021_02_09_430550.txt