Directed Scattering for Knowledge Graph-based Cellular Signaling Analysis

TL;DR

This paper introduces a novel directed scattering autoencoder framework that effectively models hierarchical and unidirectional properties of cellular signaling networks, improving graph embedding and network inference tasks.

Contribution

The paper presents the Directed Scattering Autoencoder (DSAE), a new method combining directed geometric scattering, autoencoders, and hyperbolic geometry for cellular signaling network analysis.

Findings

Outperforms existing methods in directed graph embedding

Effectively captures hierarchical structures in cellular networks

Improves cellular signaling network inference accuracy

Abstract

Directed graphs are a natural model for many phenomena, in particular scientific knowledge graphs such as molecular interaction or chemical reaction networks that define cellular signaling relationships. In these situations, source nodes typically have distinct biophysical properties from sinks. Due to their ordered and unidirectional relationships, many such networks also have hierarchical and multiscale structure. However, the majority of methods performing node- and edge-level tasks in machine learning do not take these properties into account, and thus have not been leveraged effectively for scientific tasks such as cellular signaling network inference. We propose a new framework called Directed Scattering Autoencoder (DSAE) which uses a directed version of a geometric scattering transform, combined with the non-linear dimensionality reduction properties of an autoencoder and the geometric properties of the hyperbolic space to learn latent hierarchies. We show this method outperforms numerous others on tasks such as embedding directed graphs and learning cellular signaling networks.