Background: The etiological heterogeneity of temporomandibular disorder (TMD) is reflected by the several biological processes involved, positioning it as a complex, multi-system disorder. Aims: 1. To characterize immune signatures in TMD. 2. To test their associations with clinical features such as impact pain, pain catastrophizing, and anxiety. Methods: We performed non-negative matrix factorization (NMF) using the CoGAPS package in R to identify patterns associated with biological processes (BPs) by analyzing transcriptomic data collected from blood samples of a multi-ethnic cohort of 173 individuals with TMD. Additionally, we employed a digital cytometry tool (CIBERSORTx) to estimate fractions of immune cell types from the transcriptomic data. We then examined correlations between the CoGAPS-derived patterns and immune cell type fractions, demographic variables, and clinical phenotypes. The patterns were further assessed for enrichment in immune and inflammatory BPs. We applied linear regression to understand the effects of clinical features, demographic variables, and their interactions on the patterns. Finally, we selected genes with high pattern amplitudes from each pattern and tested them for differential expression between individuals with high and low impact pain, and by different levels of pain catastrophizing and anxiety. Results: The 18 patterns derived from CoGAPS showed significant correlations with immune cell types, demographic variables, and clinical features. Gene set enrichment analysis revealed that 4 of the 18 patterns were enriched in immune and inflammatory pathways. Notably, 18 genes (including cytokines CXCL9, CXCL10, CXCL11) from one of these 4 patterns also showed higher expression in individuals with high impact pain than in those with low impact pain. Conclusion: Our study provided insight into the inflammatory context of TMD, which influences the clinical features of pain, while elucidating key underlying biological processes.