Ontology matching (OM) is one of the most challenging problems in the Semantic Web. Despite significant progress over the past few decades, evidenced by numerous approaches and tools, OM remains a complex task that often requires extensive hand-tuning for specific ontologies. Motivated by this shortcoming and inspired by recent advances in applying deep learning (DL) to graph structures, we introduce GATHER (GrAph Transformer enHancEd by Relations for OM) as an efficient solution to the existing challenges faced by OM methods that rely on Graph Convolutional Networks (GCNs). Unlike GCNs, which encounter issues such as over-smoothing and over-squashing, GATHER leverages Graph Transformers (GTs) to effectively capture complex relational dependencies within ontologies. In contrast to current embedding-based OM approaches, GATHER excels by seamlessly integrating both concept and relational features of input ontologies. It utilizes BERT-based encoder to capture semantic features, followed by the extraction of structural features using a GT that encompasses both concept and relational aspects. These semantic and structural features are combined through a gated network, facilitating effective mapping identification via an embedding matcher. Extensive experiments across multiple datasets demonstrate the superiority of GATHER over existing state-of-the-art methods, highlighting its potential to advance the field of OM.