Parkinson's disease (PD) is the most common progressive neurological disorder compromising motor functions. However, nonmotor symptoms, such as gastrointestinal (GI) dysfunction, precede those affecting movement. Evidence of an early involvement of the GI tract and enteric nervous system highlights the need for better understanding of the role of gut microbiota in GI complications in PD. Here, we investigate the gut microbiome of patients with PD using metagenomics and serum metabolomics. We integrate these data using metabolic modeling and construct an integrative correlation network giving insight into key microbial species linked with disease severity, GI dysfunction, and age of patients with PD. Functional analysis reveals an increased microbial capability to degrade mucin and host glycans in PD. Personalized community-level metabolic modeling reveals the microbial contribution to folate deficiency and hyperhomo-cysteinemia observed in patients with PD. The metabolic modeling approach could be applied to uncover gut microbial metabolic contributions to PD pathophysiology.