Properly integrating multiple modalities allows for a deeper understanding of complex biological processes. This study presents an unsupervised method of integrating proteomics, lipidomics, and metabolomics to reveal cross-omic trends, often overlooked by traditional analysis methods, in pituitary pars intermedia dysfunction (PPID). PPID is a neuroendocrine disorder in horses caused by the degeneration of hypothalamic dopaminergic neurons, leading to abnormal pituitary function. From the 6,436 detected molecular features, a traditional supervised approach found 201 significantly differentially expressed features. However, using Multi-Omics Factor Analysis (MOFA), trends not seen in the conventional approach were noted, such as divergence of C4BPA with triglycerides in the factor explaining the most lipidomic variance between the healthy and PPID-diagnosed horses, contradictory to current literary knowledge. Furthermore, using a random forest classifier, we highlighted potential biomarkers for PPID, which can be explored as potential diagnostic and therapeutic targets.