Hypergraph factorization for multi-tissue gene expression imputation

Abstract

Integrating gene expression across tissues and cell types is crucial for understanding the coordinated biological mechanisms that drive disease and characterize homoeostasis. However, traditional multi-tissue integration methods either cannot handle uncollected tissues or rely on genotype information, which is often unavailable and subject to privacy concerns. Here we present HYFA (hypergraph factorization), a parameter-efficient graph representation learning approach for joint imputation of multi-tissue and cell-type gene expression. HYFA is genotype agnostic, supports a variable number of collected tissues per individual, and imposes strong inductive biases to leverage the shared regulatory architecture of tissues and genes. In performance comparison on Genotype–Tissue Expression project data, HYFA achieves superior performance over existing methods, especially when multiple reference tissues are available. The HYFA-imputed dataset can be used to identify replicable regulatory genetic variations (expression quantitative trait loci), with substantial gains over the original incomplete dataset. HYFA can accelerate the effective and scalable integration of tissue and cell-type transcriptome biorepositories.

Transcriptome‐Wide Association Studies (TWAS): Methodologies, Applications, and Challenges

“…Larger eQTL datasets are available but only for select tissues, such as brain (PsychENCODE) (PsychENCODE Consortium et al, 2015) and whole blood (eQTLGen Consortium) (Võsa et al, 2021). One approach to expand the scale of a transcriptome dataset is Hypergraph Factorization (Viñas et al, 2023). The neural network approach allows for the imputation of gene expression in hard-to-collect tissues using available expression data in more accessible tissues, e.g., whole blood.…”

SAF: Smart Aggregation Framework for Revealing Atoms Importance Rank and Improving Prediction Rates in Drug Discovery

“…The second approach includes attention schemes based on GAT. , The concept of GAT forces the model to decide to which edge it should attend to. If the prediction performance improves, then this implies that the selected subset of edges is important to the objective.…”

Analysis of higher order interactions quantifies co-ordination in the epigenome and reveals novel biological relationships in Kabuki syndrome

“…They have been used to examine early human embryology 14,15 and hypergraph topology has been used to assess mechanism in neural, ecological and social systems 11 . Recently, efficient imputation of multi-tissue and cell-type gene expression has been achieved using a hypergraph approach 16 . However, hypergraphs have not yet been used to assess mechanistic relationships in human diseases.…”