Guoliang Deng, Tong Cui, Yifeng Xu, Wanning Gao, Nan Li, Lijuan Jiang, Wenfeng Zhang
Background/Objectives: Oral Xuesaitong soft capsules have been reported to improve functional outcomes after ischemic stroke, but their molecular mediators remain uncertain. We evaluated human genetic support for pharmacology-informed Panax notoginseng saponin-related candidate genes. Methods: A targeted pharmacological audit was followed by cis-QTL Mendelian randomization (MR) and regional colocalization using GISCOME outcomes. The severity-adjusted ordinal modified Rankin Scale (mRS) was primary. Eleven CKLF, KMO, and MT2A gene-tissue hypotheses formed the eQTL testing family; HMOX1 and OGA underwent pQTL analyses. A 13-test Bonferroni sensitivity calculation combined these hypotheses. Colocalization required PP.H4 ≥ 0.80. KMO analyses covered 15 GTEx tissues, three outcomes, and two external eQTL datasets. The summary-statistics study was not prospectively registered. Results: The audit identified 29 component-protein relations involving 19 proteins. Further, 7 of 11 eQTL hypotheses were harmonized. Higher genetically predicted KMO expression in whole blood was associated with better ordinal mRS (β = -0.906; 95% CI, -1.521 to -0.290; p = 0.0039; adjusted p = 0.043 for 11 tests and 0.051 for 13 tests). HMOX1 and OGA MR estimates were nonsignificant. Primary PP.H4 values were 0.033, 0.027, and 0.535 for HMOX1, OGA, and KMO, respectively. None of the 45 GTEx or six external comparisons met the criterion under primary priors. Conclusions: No candidate met the joint MR and colocalization criterion. KMO significance was sensitive to multiplicity and standard-error specification, while colocalization support was insufficient and prior-sensitive. These findings do not establish a therapeutic target or test the efficacy of oral Xuesaitong.