Genome-wide association studies revealed single nucleotide polymorphisms (SNPs) associated with inflammatory bowel disease (IBD) in coding regions of autophagy-related genes, such as ATG16L1 and NOD2. These variants contribute to IBD pathogenesis by disrupting autophagy and impairing its key function to maintain intestinal homeostasis, regulating immune responses and the removal of intracellular pathogens. However, the majority of IBD-associated SNPs occur in non-coding genomic regions, and their impact on key pathways such as autophagy is currently unknown. In this work, we address this by evaluating the impact of non-coding SNPs on autophagy using our state-of-the-art in silico pipelines.
We analysed 1695 genotyped Crohn’s disease (CD) and 941 ulcerative colitis (UC) patients individually through a novel workflow that integrated two computational resources we recently developed: the iSNP precision medicine pipeline which allows patient-specific genome and cellular signaling analysis, and AutophagyNet, a comprehensive database of autophagy core genes and regulatory molecules.
Our analysis revealed that IBD patients cluster into distinct groups based on the predicted impact of non-coding SNPs on autophagy core genes and regulatory proteins. While autophagy dysfunction in CD is traditionally attributed to an ATG16L1 coding risk variant, we found that 2 non-coding SNPs can indirectly perturb ATG16L1 and other autophagy proteins. We also found other non-coding SNPs perturbing several known autophagy regulator proteins in both UC and CD. These include CDC37 which we previously identified as key in autophagy dynamics, and TWIST1 which regulates LC3-associated phagocytosis. In UC, SNP-affected autophagy regulating proteins were different between patient clusters, whilst core autophagy genes regulated by these proteins remained largely the same in all clusters. This indicates different pathomechanistic routes to reach the same effect. However, in CD core autophagy genes perturbed by SNP-affected regulators were different in various clusters, pointing out different autophagy modulations in patient groups. Interestingly, we found no coding SNP-perturbed core autophagy genes in 328 UC patients (34.8%) and 653 CD patients (38.5%). However, autophagy regulating proteins were still perturbed in all these patients, showing the significance of examining non-coding regulation of autophagy. We observed that ~50% of SNP-affected regulatory proteins overlapped between UC and CD patients, revealing significant differences in the regulatory mechanisms by which autophagy is impaired.
Overall, our findings provide novel insights into heterogenetic mechanisms by which autophagy is perturbed in IBD patients. By unravelling patient-specific impacts of IBD-associated non-coding SNPs on autophagy, we lay the foundations for advancing precision medicine efforts in IBD.

Figure 1 - Schematic overview of study workflow.
Figure 2 - Clustering of UC (right) and CD (left) patients based on affected autophagy core members and regulators. Core proteins with coding SNPs are highlighted green on the Y axis. The upper column colorbar shows patient clusters from a different cluster calculation method, while the lower column colorbar shows in dark blue those patients where autophagy is affected by both coding and non-coding SNPs, while light blue indicates those patients, where only regulatory SNPs are present.