Fig. 2

Study design. First, using MR to identify potential causal proteins for LUAD by utilizing plasma pQTL data from Ferkingstad’s study and GWAS data from the TRICL Consortium. Otherwise, we replicated the primary analysis in dependent cohorts, using plasma pQTL data from Sun’s study and LUAD GWAS data from FinnGen Cohort. Second, sensitivity analyses were used to validate our primary findings, including Cochran Q test, MR-Egger test, and MR-PRESSO test. Bidirectional MR analysis and Steiger filtering were used to ensure the directionality of causality. Phenotype scanning and Bayesian co-localization were employed to detect potential horizontal pleiotropy. Third, we conducted mediation analysis using a two-step MR for proteins displaying potential causality with both LUAD and risk factors. Fourth, we mapped the interaction network among the identified proteins and their associations with the targets of current LUAD drugs. Lastly, we searched for an updated list of druggable genes, the ChEMBL database and a clinical trials registry website to evaluate the druggability of the identified proteins. ^#LUAD Risk factors: age of smoking initiation, number of cigarettes smoked daily, pack years of smoking, maternal smoking around birth, any parental history of lung cancer, average weekly beer intake, leisure activities, and chronic obstructive pulmonary disease (COPD). LUAD: lung adenocarcinoma; GWAS: genome-wide association study; TRICL: Transdisciplinary Research in Cancer of the Lung; pQTLs: protein quantitative trait locus; LD: linkage disequilibrium; SNPs: single nucleotide polymorphisms; FDR: false discovery rate; MR: Mendelian Randomization; MR-PRESSO: MR Pleiotropy Residual Sum and Outlier; PP.H4: posterior probability of H4