id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
10_1101-2020_11_17_386649	Danciu, Daniel	Topology-based Sparsification of Graph Annotations	2021	15	.pdf	application/pdf	8205	774	67	Experiments on 10,000 RNA-seq datasets show that RowDiff combined with MultiBRWT results in a 30% reduction in annotation footprint over Mantis-MST, the previously known most a binary matrix, where the k-mer set indexes the rows and each annotation label specifies a column. Starting from any vertex in the de Bruijn graph, Algorithm 1 defines a traversal leading to an anchor Each row in a RowDiff-transformed annotation matrix has the same or fewer set bits than A naı̈ve implementation of the RowDiff construction would be to load the matrix A in memory, and gradually replace its rows with their sparsified counterpart, while traversing the graph. We now note that, when querying annotations for paths in the graph, or sets of rows corresponding to vertices We constructed annotated de Bruijn graphs from the RNA-Seq data set in the same We now compare the representation size for RowDiff and other state-of-the-art graph annotation compression methods.	./cache/10_1101-2020_11_17_386649.pdf	./txt/10_1101-2020_11_17_386649.txt