It is suggested that subtyping of complex inflammatory diseases can be based on genetic susceptibility and relevant environmental exposure (G+E). We propose that using matched cellular phenotypes in human subjects and corresponding preclinical models with the same G+E combinations is useful to this end. As an example, defective Paneth cells can subtype Crohn's disease (CD) subjects; Paneth cell defects have been linked to multiple CD susceptibility genes and are associated with poor outcome. We hypothesized that CD susceptibility genes interact with cigarette smoking, a major CD environmental risk factor, to trigger Paneth cell defects. We found that both CD subjects and mice with ATG16L1T300A (T300A; a prevalent CD susceptibility allele) developed Paneth cell defects triggered by tobacco smoke. Transcriptional analysis of full-thickness ileum and Paneth cell-enriched crypt base cells showed the T300A-smoking combination altered distinct pathways, including proapoptosis, metabolic dysregulation, and selective downregulation of the PPARγ pathway. Pharmacologic intervention by either apoptosis inhibitor or PPARγ agonist rosiglitazone prevented smoking-induced crypt apoptosis and Paneth cell defects in T300A mice and mice with conditional Paneth cell-specific knockout of Atg16l1. This study demonstrates how explicit G+E can drive disease-relevant phenotype and provides rational strategies for identifying actionable targets.