2022 # scSGL: kernelized signed graph learning for single-cell gene regulatory network inference

**Abstract:** Motivation
Elucidating the topology of gene regulatory networks (GRNs) from large single-cell RNA sequencing (scRNAseq) datasets, while effectively capturing its inherent cell-cycle heterogeneity and dropouts, is currently one of the most pressing problems in computational systems biology. Recently, graph learning (GL) approaches based on graph signal processing (GSP) have been developed to infer graph topology from signals defined on graphs. However, existing GL methods are not suitable for …

Help me understand this report

Search citation statements

Paper Sections

Select...

3

2

Citation Types

0

11

0

Year Published

2023

2023

Publication Types

Select...

2

2

1

Relationship

1

4

Authors

Journals

(11 citation statements)

0

11

0

“…In [ 31 ], ( 2 ) is extended to learn an unknown signed graph G based on the assumption that the graph signals admit (i) low-frequency (smooth) representation over , and (ii) high-frequency (nonsmooth) representation over . Smoothness and non-smoothness of the graph signals with respect to signed graphs are defined as follows: (1) Signal values on nodes that are connected with positive edges are similar to each other; (2) Signal values on nodes that are connected with negative edges are dissimilar to each other.…”

confidence: 99%

“…In [ 31 ], ( 2 ) is extended to learn an unknown signed graph G based on the assumption that the graph signals admit (i) low-frequency (smooth) representation over , and (ii) high-frequency (nonsmooth) representation over . Smoothness and non-smoothness of the graph signals with respect to signed graphs are defined as follows: (1) Signal values on nodes that are connected with positive edges are similar to each other; (2) Signal values on nodes that are connected with negative edges are dissimilar to each other.…”

confidence: 99%

“…On the other hand, if i and j are connected by an inhibitory edge, their expressions should be dissimilar, i.e., high frequency. These assumptions are biologically reasonable and have been validated in [ 31 ]. Based on these assumptions, the signed graph G is learned by minimizing with respect to and maximizing with respect to .…”

confidence: 99%