Page 164 - EJMO-9-2

Eurasian Journal of
Medicine and Oncology Genetic insights into CAD drug targets

extensive phenotypic data, which helps in evaluating molecular docking was employed to assess binding
the performance of potential drug targets across various energy and interaction patterns between candidate
phenotypes. In this study, the PheWAS dataset included drugs and targets at the atomic level. Molecular docking
more than 15,000 binary phenotypes and 1,500 continuous simulations allowed us to evaluate the binding affinity
phenotypes derived from individuals in the UK Biobank, and interaction characteristics between ligands and
including a subset with exome sequencing. targets, thereby prioritizing those with high binding
By analyzing these data, we systematically evaluated the affinity and favorable interaction patterns for further
role of candidate targets across different populations and experimental validation and optimization. In this study,
phenotypes, thereby identifying potential side effects and molecular docking analysis of protein-ligand interactions
pleiotropic features. Standard statistical analysis methods was conducted using MOE 2019 software to simulate the
were applied to detect associations between candidate gene binding process between candidate drugs and proteins
variants and multiple phenotypes, controlling for possible encoded by the target genes. The three-dimensional
confounding factors. This comprehensive PheWAS structure data of the drugs were obtained from the
analysis provided an in-depth understanding of the PubChem Compound Database (https://pubchem.ncbi.
impact of genetic factors on complex phenotypes, offering nlm.nih.gov/). The three-dimensional structure data of
scientific evidence for the efficacy and safety of candidate the proteins were retrieved from the Protein Data Bank
drug targets. (PDB; http://www.rcsb.org/). The research design process
is displayed in Figure 1.
2.9. Enrichment analysis
To explore the functional characteristics and biological 3. Results
relevance of druggable genes, gene ontology (GO) and 3.1. Genes causally associated with CA risk during
Kyoto Encyclopedia of Genes and Genomes (KEGG) the discovery phase
enrichment analyses were performed using the R package
“clusterProfiler.” The GO enrichment analysis included In the discovery phase, we conducted MR analysis on
18
three aspects: BPs, molecular function (MF), and cellular atherosclerotic patients. The cohort study was sourced
component (CC), which describe the functions of genes from the GWAS catalog database, including 16,041 cases
and the BPs they are involved in. KEGG analysis provided and 440,307 controls. Using the IVW method and after
information on metabolic pathways, helping to elucidate FDR correction (FDR<0.05), 130 genes were identified as
the specific functions of genes within biological systems significantly associated with the risk of atherosclerosis. All
and their roles in metabolic networks. significant IVs and the complete results of the MR analysis
are summarized in Table S5.
2.10. Protein-protein interaction (PPI) network
construction 3.2. Replication phase 9 genes remain significant in
To better understand whether a protein interacts with independent CA cohorts
another protein within the cell, a PPI network was In the replication phase, GWAS data from the Finnish
constructed using STRING, and the network was further FinnGen database (including 8,279 cases and 261,098
19
visualized using Cytoscape (V3.9.1). In addition, European-ancestry controls) and the UK Biobank database
GeneMANIA was employed to study protein interactions. (including 14,334 cases and 346,860 European-ancestry
2.11. Candidate drug prediction controls) were used. MR analysis was conducted similarly
to the discovery phase. Using the IVW method, 131 gene
Assessing protein-drug interactions is crucial for expressions and 78 genes (all of which passed heterogeneity
determining the potential of target genes as drug targets. and horizontal pleiotropy tests) were found to have a causal
In this study, the drug signatures database (DSigDB; relationship with CA risk (Tables S6 and S7). Subsequently,
http://dsigdb.tanlab.org/DSigDBv1.0/) was used to analyze a cross-analysis of potential drug targets identified across
associations between drugs and compounds and the target the three databases resulted in the identification of nine
genes. By inputting target genes, researchers were able to unique potential druggable genes for CA (Figure 2).
20
identify drugs or compounds significantly associated with
these genes (Tables S2-S5, S21). For the nine unique potential drug genes, MR analysis
using cis-eQTLs from 49 tissues (GTEx V.8 version)
2.12. Molecular docking suggested differences for CHD4, tumor necrosis factor
To further explore the effect of candidate drugs on (TNF), FKRP, CD164L2, DHX36, lipoprotein lipase (LPL),
target genes and evaluate the druggability of these genes, FES, and TAF1A across tissues (p<0.05) (Tables S8-S10).

Volume 9 Issue 2 (2025) 156 doi: 10.36922/ejmo.7387