Supplementary Material _ Oliveira, CS et al.

KEGG Pathway Database (Kanehisa, 2019; Kanehisa et al., 2025; Kanehisa &amp; Goto, 2000) was used to collect genes on insulin signaling, lipid metabolism and inflammation pathways, underlying metabolic disorders common to Alzheimer's Disease and obesity. The prefix “hsa” was selected for human data, and three search terms were inserted: “inflammation”, “insulin” or “lipid”. The pathways with general descriptions surrounding the search terms were selected <b>(Supplementary Material 1)</b>, and the section “Gene” was exported manually. Pathways with descriptions surrounding specific diseases or molecules were excluded, as well as duplicate genes, resulting in a total of 1879 genes <b>(Supplementary Material 2).</b>A list of genes constituting the network of insulin signaling, lipid metabolism and inflammation pathways was used as entries in the miRWalk database (Sticht et al., 2018) to obtain all miRNAs predicted to interact with the sequences of these genes. The prediction score of 1.0 (highest confidence) was set as a filter, resulting in a total of 1786 miRNA sequences <b>(Supplementary Material 3)</b>.To cover all miRNAs likely to be involved in metabolic disorders common to AD and obesity, another database was accessed, in which it is possible to insert search terms. Thus, The Human microRNA Disease Database (HMDD) v4.0 (Cui et al., 2024) was used to access miRNA sequences via disease descriptors (‘obesity’ and ‘Alzheimer’s disease’), allowing the identification of miRNAs related to obesity and Alzheimer’s disease in a general manner. There were 422 miRNAs related to obesity and 779 to Alzheimer's disease; these two lists were overlapped, and we were able to identify 113 of them that were common between both conditions <b>(Supplementary Material 4).</b>Common miRNAs between the list from miRWalk and the list from HMDD were used for the following analysis. We found 107 miRNAs common to these two lists <b>(Supplementary Material 5)</b>, indicating miRNAs associated both to the molecular pathways and the diseases in the study.In order to satisfactorily address the miRNAs-target gene relationships, two cutoff parameters were adopted through centrality analysis to obtain the most relevant target genes. Only genes with betweenness and degree values higher than the third quartile of both parameters were considered highly central to the network, and chosen to build the GRNs (n=206 — target genes with high centrality) <b>(Supplementary Material 6)</b>.Finally, Over-representation Analysis (ORA) was performed using KEGG and Reactome databases. Common genes indicated in categories of both analysis were identified (n=30 with p ≤ 0.00001 in KEGG and Reactome databases) <b>(</b><b>Supplementary</b><b> Marerial 7).</b>