Implications of Topological Imbalance for Representation Learning on Biomedical Knowledge Graphs

TL;DR

This paper investigates how topological imbalances in biomedical knowledge graphs influence embedding-based predictions, revealing that densely connected entities tend to be overrepresented regardless of biological relevance.

Contribution

It uncovers the impact of structural biases in knowledge graphs on model outputs and emphasizes the importance of careful data modeling in biomedical applications.

Findings

Densely-connected entities are often highly ranked regardless of biological relevance.

Graph perturbation experiments show models are influenced more by entity frequency than biological information.

Structural imbalances can bias predictions, affecting their biological interpretability.

Abstract

Adoption of recently developed methods from machine learning has given rise to creation of drug-discovery knowledge graphs (KG) that utilize the interconnected nature of the domain. Graph-based modelling of the data, combined with KG embedding (KGE) methods, are promising as they provide a more intuitive representation and are suitable for inference tasks such as predicting missing links. One common application is to produce ranked lists of genes for a given disease, where the rank is based on the perceived likelihood of association between the gene and the disease. It is thus critical that these predictions are not only pertinent but also biologically meaningful. However, KGs can be biased either directly due to the underlying data sources that are integrated or due to modeling choices in the construction of the graph, one consequence of which is that certain entities can get topologically overrepresented. We demonstrate the effect of these inherent structural imbalances, resulting in densely-connected entities being highly ranked no matter the context. We provide support for this observation across different datasets, models as well as predictive tasks. Further, we present various graph perturbation experiments which yield more support to the observation that KGE models can be more influenced by the frequency of entities rather than any biological information encoded within the relations. Our results highlight the importance of data modeling choices, and emphasizes the need for practitioners to be mindful of these issues when interpreting model outputs and during KG composition.