Mendelian Randomization Unveils Poly(ADP-ribose) Polymerase 1 as Causal Plasma Protein Target in Melanoma, Integrated Multi-omics Authors: Feng Lu, Wenkang Luan, Hanyi Jiang, Shaojun Ma, Hongru Ruan Published: 2026 ABSTRACT Background:The molecular regulators driving melanoma pathogenesis remain incompletely defined. Although plasma proteins show associations with melanoma, their causal relationships are unclear.Methods:We performed Mendelian randomization (MR) using proteomic data from the UK Biobank Pharma Proteomics Project and deCODE genetics to identify plasma proteins causally linked to melanoma. Colocalization, summary data–based MR, heterogeneity in dependent instruments, and MR-phenome-wide association study analyses identified druggable targets. Transcriptome profiling of melanoma tissues and single-cell functional validation assessed the target roles.Results:Poly(adenosine diphosphate-ribose) polymerase 1 (PARP1) was established as a causal risk factor and druggable target for melanoma. Elevated PARP1 expression in melanoma tissues correlated with poor patient survival. Functional analyses revealed that PARP1 coordinates protumorigenic functions with cell-specific roles.Conclusions:In conclusion, our results demonstrate that PARP1 is a key driver of melanoma pathogenesis and represents a potential specific molecular target for therapeutic intervention. Takeaways Question: Which plasma proteins have a causal, rather than merely associative, relationship with melanoma and may represent potential druggable targets? Findings: Using Mendelian randomization on large-scale proteomic data, we identified 17 candidate proteins. Colocalization robustly confirmed poly(adenosine diphosphate-ribose) polymerase 1 (PARP1) as a key causal risk factor. Multi-omics validation showed that elevated PARP1 expression in melanoma tissues was correlated with poor survival, and single-cell analysis linked it to protumorigenic functions. Meaning: This study provided genetic evidence that the plasma protein PARP1 is a causal driver and a promising therapeutic target for cutaneous melanoma. INTRODUCTION Cutaneous melanoma represents the most aggressive and lethal subtype of skin malignancies, accounting for approximately 90% of skin cancer–related deaths worldwide.1,2 Beyond the common cutaneous form, melanoma can also arise in mucosal and central nervous system sites, where delayed diagnosis and enhanced invasiveness contribute to poor clinical outcomes.3,4 Nodular melanoma, characterized by rapid vertical growth and high metastatic potential even at low Breslow thickness, underscores the urgent need to identify precise molecular biomarkers to identify early aggressive phenotypes.5,6 Although advances in screening and diagnostic techniques have improved early detection, the growing concern of overdiagnosis highlights the importance of establishing biomarker-based precision diagnostic systems for melanoma.7,8 Plasma proteins play multifaceted roles in melanoma development and progression by modulating the tumor microenvironment and regulating signaling pathways involved in cell proliferation, migration, and immune evasion. As liquid biopsy–accessible biomarkers, plasma proteins hold great promise for early detection, prognostic assessment, and therapeutic monitoring.9–11 However, most existing studies are limited by small sample sizes and observational designs, making it difficult to determine causal relationships between plasma protein alterations and melanoma risk. Mendelian randomization (MR), leveraging the natural randomization of genetic variants, offers a powerful approach to infer causality while minimizing confounding effects.12 To address this gap, we used an MR analysis leveraging plasma protein quantitative trait loci (pQTLs) from the UK Biobank Pharma Proteomics Project (UKB-PPP) and deCODE genetics databases to systematically identify candidate causal plasma proteins for melanoma pathogenesis. Following this initial screening, a stringent, multistep validation pipeline—including colocalization, summary data–based MR (SMR), heterogeneity in dependent instruments (HEIDI), and MR-phenome-wide association study (MR-PheWAS) analyses—was applied to robustly confirm causal relationships and assess the druggable potential of prioritized proteins. Through this integrated approach, poly(adenosine diphosphate-ribose) polymerase 1 (PARP1) was established as a melanoma risk factor with therapeutic targeting potential. PARP1 is a nuclear enzyme essential for single-strand DNA break repair and genomic stability maintenance.13 Beyond DNA repair, PARP1 modulates chromatin remodeling, transcriptional regulation, apoptosis, and inflammatory signaling, influencing tumor microenvironment dynamics through pathways such as nuclear factor kappa B.14 Although mutations in PARP1 are rare, elevated PARP1 expression has been implicated in promoting melanoma cell survival, cooperating with BRAF mutations to drive tumor progression, and modulating immune responses, thereby increasing melanoma susceptibility.15 As a therapeutic target, PARP1 inhibitors exploit the “synthetic lethality” mechanism to selectively eliminate tumor cells with defective DNA repair and may synergize with immunotherapy to enhance treatment efficacy.16 Currently, PARP inhibitors have entered early-phase clinical trials for melanoma.17 Genomic studies suggest that a subset of melanomas may be susceptible, with an estimated 20%–40% of cutaneous tumors harboring relevant DNA repair gene alterations, which are associated with preclinical sensitivity to PARP inhibition.18 However, robust phase III efficacy data and a clinically validated biomarker definition are still lacking.19 This gap hinders the translation of PARP1 biology into precision therapy for melanoma. To address this, we applied an integrated genetic framework to systematically identify and validate causal plasma proteins for melanoma risk. We then focused on evaluating the biological and therapeutic relevance of the top-prioritized protein, PARP1. Our findings provide causal evidence to inform future biomarker and therapeutic strategies in melanoma. MATERIALS AND METHODS Study Design Description We performed a 2-sample MR analysis to assess causal relationships between plasma proteins and melanoma. Cis-acting pQTLs from the UKB-PPP database (2940 proteins) and deCODE database (4907 proteins) served as instrumental variables (IVs) for exposures, with melanoma as the outcome. Subsequent analyses included colocalization, SMR, HEIDI, and MR-PheWAS analyses. To validate the role of key plasma proteins in melanoma, we integrated bulk RNA-seq data from melanoma tissues in The Cancer Genome Atlas (TCGA) and normal skin tissues from Genotype-Tissue Expression (GTEx), along with single-cell transcriptomes from the CancerSEA database (http://biocc.hrbmu.edu.cn/CancerSEA). This study does not involve additional ethical approval, as it is based on the public dataset. GWAS Data Genome-wide association study (GWAS) data for 2940 plasma proteins were obtained from a large-scale analysis of plasma proteomics and human health in the UK Biobank.20 This study characterized plasma proteomic profiles of 54,219 Europeans and generated corresponding GWAS datasets, publicly available through the UKB-PPP database (https://www.synapse.org/Synapse:syn51364943/wiki/622119) (Table 1). Table 1.Details of the GWAS Summary Data Included in the StudyNameSitePopulationSample SizeWeb Source2940 plasma proteinsUKB-PPPEuropean54,219 https://www.synapse.org/Synapse:syn51364943/wiki/622119 4907 plasma proteinsdeCODE geneticsEuropean35,559 https://www.decode.com/summarydata Skin melanomaFinnGen biobankEuropean384,502 https://r12.finngen.fi/pheno/C3_MELANOMA_SKIN_EXALLC Details of the GWAS Summary Data Included in the Study GWAS data for 4907 plasma proteins were sourced from a comprehensive analysis linking plasma protein levels to 373 diseases and traits.21 This research characterized plasma proteomic profiles of 35,559 Icelanders, with the resulting aptamer-based GWAS datasets publicly available in the deCODE genetics database (https://www.decode.com/summarydata/) (Table 1). Melanoma GWAS data (finngen_R12_C3_MELANOMA_SKIN_EXALLC) were sourced from the FinnGen biobank.22 This dataset comprises 5753 cases with skin melanoma and 378,749 European-ancestry controls (Table 1). Selection of IVs for MR For valid IV selection in MR, 3 core assumptions must be satisfied: association with the exposure, the independence assumption, and exclusive restriction to the outcome through the exposure. The screening criteria applied were GWAS significance, a P value of less than 5 × 10−8 for all candidate single-nucleotide polymorphism (SNPs); linkage disequilibrium (LD) pruning, exclusion of SNPs in LD (window: 10,000 kb; r2 threshold > 0.001); instrument strength, removal of SNPs with F-statistic values of less than 10; cis-regulatory window, restriction to variants within ±1000 kb of protein-coding regions. MR Causal Effect Estimation During harmonizing outcome and exposure data, palindromic SNPs exhibiting intermediate allele frequencies were excluded.23 Causal relationships between plasma proteins and melanoma were evaluated using the inverse-variance weighting (IVW) framework, weighted median, MR-Egger test, and weighted mode methods when multiple IVs remained; otherwise, the Wald ratio was used for single IVs. The IVW framework served as the primary analysis method in the absence of pleiotropy, as it assumes that instruments influence outcomes exclusively through the exposure pathway,23,24 with other methods providing supplementary robustness checks. Sensitivity analyses encompassed heterogeneity assessment via Cochran Q statistic in both IVW and MR-Egger frameworks, pleiotropy evaluation using MR-Egger intercept test, and leave-one-out validation to confirm that estimates were not disproportionately influenced by individual SNPs. Colocalization Analysis Colocalization analysis was performed using the “COLOC” R package to investigate whether melanoma and plasma proteins share causal genetic variants, specifically for proteins significantly associated with melanoma.25 The evidence was quantified using posterior probabilities for 5 competing hypotheses (H0–H4). These posterior probabilities are denoted as PPH0–PPH4: PPH0 indicates no association, PPH1 signifies association with plasma proteins only, PPH2 denotes association with melanoma only, PPH3 represents independent associations with both traits, and PPH4 implies shared causal variation.25 A PPH4 value of 0.8 or higher defined strong colocalization, classifying the involved genes as potential drug targets.25 SMR and HEIDI Test SMR and HEIDI tests were used to differentiate between pleiotropic effects and genetic linkage models.26 As a supplementary approach, SMR analysis evaluated causal relationships between melanoma and both expression quantitative trait loci (eQTLs) and pQTLs, with eQTL data obtained from eqtlGEN (https://www.eqtlgen.org/), and a P value of less than 0.05 defined as significant correlation.27 The HEIDI test further confirmed that associations of plasma proteins with melanoma (P > 0.05) stemmed from shared genetic variations rather than LD.27 MR-PheWAS Analysis To investigate potential side effects of druggable targets, we performed MR-PheWAS analysis using plasma proteins as exposures and summary data from 2469 diseases in the FinnGen R12 cohort as outcomes. MR analysis was conducted via the Wald ratio or IVW method, with causal effects considered statistically significant at a P value of less than 5 × 10−8. This approach provides insights into complex genetic features and facilitates the evaluation of safety and efficacy of druggable targets. Public Human Tissue Samples and Analysis RNA expression data from 461 melanoma tissues in TCGA and 558 normal skin tissues in GTEx were analyzed via the GEPIA platform (http://gepia.cancer-pku.cn/index.html). Additionally, prognostic data from the TCGA-melanoma cohort were assessed through GEPIA to examine whether PARP1 expression levels influence patient survival. To evaluate the functional role of PARP1 in melanoma cells, RNA-seq data from 307 single cells—cultured from biopsies of 3 distinct melanoma patients—were analyzed using the CancerSEA database (http://biocc.hrbmu.edu.cn/CancerSEA). ELISA of Clinical Plasma Specimens Clinical plasma samples from 5 melanoma patients and 5 healthy controls were collected from the Affiliated People’s Hospital of Jiangsu University. The study was approved by the human research ethics committee of the Affiliated People’s Hospital of Jiangsu University. PARP1 levels were measured using a human PARP ELISA kit (Cusabio, Wuhan, China). Following the manufacturer’s protocol, samples and standards were added to precoated plates, incubated with the supplied detection antibody and horseradish peroxidase‑conjugated secondary antibody. After further incubation, washing, and addition of tetramethylbenzidine substrate, the reaction was stopped with sulfuric acid, and absorbance was read at 450 nm. Sample concentrations were determined from the standard curve. RESULTS Detecting Plasma Proteins Exerting Causal Influence on Melanoma We first investigated the causal relationship between plasma proteins and melanoma through MR analysis. To enhance result reliability and universality, we analyzed 2 independent plasma protein GWAS datasets: the UKB-PPP database covering 2940 plasma proteins, and the deCODE genetics database comprising 4907 plasma proteins. According to the screening criteria detailed in the Materials and Methods section, 5620 SNPs and 5502 SNPs were, respectively, identified as IVs for MR analysis, corresponding to 2940 plasma proteins from the UKB-PPP database and 4907 proteins from the deCODE genetics database. Comprehensive IV details are provided in Supplemental Digital Contents 1 and 2. (See spreadsheet, Supplemental Digital Content 1, which displays how 5620 SNPs related to 2940 plasma proteins were included as IVs [UKB-PPP], https://links.lww.com/PRSGO/F222.) (See spreadsheet, Supplemental Digital Content 2, which displays how 5502 SNPs related to 4907 plasma proteins were included as IVs [deCODE genetics], https://links.lww.com/PRSGO/F223.) As shown in Figure 1, analyses of the UKB-PPP database identified 122 plasma proteins exhibiting causal relationships with melanoma through the Wald ratio or IVW methods, with 59 demonstrating protective effects and 63 exhibiting hazardous effects. Identifying plasma proteins that exert causal effects in melanoma from the UKB-PPP database. Forest plot showing the odds ratio and 95% confidence interval of the 122 plasma proteins that were found to have a causal relationship with melanoma through the Wald ratio or IVW methods. The data of 2940 plasma proteins were extracted from the UKB-PPP database. Analysis of the deCODE genetics database revealed 96 plasma proteins causally influencing melanoma risk, split equally between 48 protective and 48 hazardous proteins (Fig. 2A). Further analysis revealed 17 plasma proteins causally linked to melanoma through the intersection of the 2 database results, specifically SFRP4, PPP1R14A, TIGIT, TNFAIP8, LDLRAP1, PECR, GRN, POSTN, DSCAM, ENO2, TIMP2, PARP1, CCL3, IL10RB, KIR2DL3, FN1, and DPEP1 (Fig. 2B). Detecting plasma proteins exerting causal influence on melanoma from the deCODE genetics database. A, Forest plot showing the odds ratio and 95% confidence interval of the 96 plasma proteins that were found to have a causal relationship with melanoma; the data of 4907 plasma proteins were taken from the deCODE genetics database. B, The intersection of the 17 significant plasma proteins from the 2 databases. PARP1 Is Identified as a Key Causal Risk Factor for Melanoma via Colocalization Analysis We performed colocalization analysis to evaluate the probability of shared causal genetic variants between pQTLs of the 17 candidate plasma proteins and melanoma-associated SNPs. Among these proteins, only PARP1 exhibited significant colocalization evidence (PPH4 > 0.8) in both the UKB-PPP (PPH4 = 0.950036) and deCODE genetics datasets (PPH4 = 0.96501) (Fig. 3A, Table 2). Regional association plots for the colocalized signals of PARP1 pQTLs and melanoma risk loci are shown in Figures 3B and C. These results indicate a high posterior probability that PARP1 pQTLs and melanoma susceptibility share the same causal genetic variation at this locus. Combined with the colocalization evidence, MR analysis robustly supports PARP1 as a causal risk factor for melanoma (Figs. 1, 2A). Table 2.Colocalization Analysis of Plasma Proteins and MelanomaPlasma ProteinPPH4 (UKB-PPP)PPH4 (deCODE)SFRP40.0613130.041263PPP1R14A0.0313340.031744TIGIT0.0888570.123112TNFAIP80.1535960.127708CCL30.0356910.071612IL10RB0.0074840.081029LDLRAP10.3093470.244572PECR0.632050.552255GRN0.0623540.061877POSTN0.0347620.038032DSCAM1.29 × 10−62.28 × 10−6ENO20.0520070.049287KIR2DL30.0204630.069984TIMP20.0342860.112543FN10.178210.065715PARP10.9500360.96501DPEP17.70 × 10−280.00088 Colocalization Analysis of Plasma Proteins and Melanoma PARP1 is identified as a key causal risk factor for melanoma via colocalization analysis. A, The PPH4 value of the colocalization analysis between the 17 significant plasma proteins and melanoma in both the UKB-PPP and deCODE databases. B, Regional plot of the colocalization of PARP1 and melanoma in the UKB-PPP database. C, Regional plot of the colocalization of PARP1 and melanoma in the deCODE database. PARP1 Can Be Used as a Potential Druggable Target for Melanoma We further applied SMR and HEIDI analyses to evaluate the potential of PARP1 as a drug target for melanoma using eQTL and pQTL data. The pQTLs for PARP1 passed both the SMR (P < 0.05) and HEIDI (P > 0.05) tests. In contrast, eQTL data showed nonsignificant SMR results (P = 0.7985) and failed the HEIDI test (P = 0.004272) (Fig. 4A, Table 3). The associated pQTL locus and effect are visualized in Figures 4B and C. These results indicate that PARP1 protein represents a druggable target for melanoma. Furthermore, MR-PheWAS analysis using the FinnGen R12 database revealed no significant genome-wide associations between PARP1 and other phenotypes, except for myeloproliferative disorders (P < 5 × 10−8) (Fig. 4D). (See spreadsheet, Supplemental Digital Content 3, which displays the MR-PheWAS results, https://links.lww.com/PRSGO/F224.) This suggests a minimal risk of adverse effects or unexpected pleiotropy if PARP1 is used as a potential therapeutic target for melanoma. Table 3.SMR and HEIDI AnalysesP_SMRP_HEIDIeQTLs of PARP10.79850.004272pQTLs of PARP10.000320.050683 SMR and HEIDI Analyses PARP1 can be used as a druggable target for melanoma. A, The P value of the SMR and HEIDI analyses based on the eQTL and pQTL data of PARP1. B, SMR locus plots illustrating the correlations between PARP1 pQTLs and melanoma risk. C, SMR effect plots demonstrating the correlations of PARP1 pQTLs with melanoma susceptibility. D, Manhattan plot showing the phenome-wide MR results of PARP1. The ordinate represents the P value in phenome-wide MR results. A dot represents a disease trait. CML, chronic myeloid leukemia. Elevated PARP1 Expression in Melanoma Tissues Correlates With Poor Overall Survival in Patients To further investigate the role of PARP1 in melanoma, we analyzed expression levels of PARP1 using public datasets (TCGA melanoma and GTEx normal skin) through GEPIA (http://gepia.cancer-pku.cn/index.html). Comparison of 461 melanoma and 558 normal tissues revealed significantly higher PARP1 expression in tumors (98% of melanoma samples expressed PARP1 at levels exceeding the mean value of the normal group) (Fig. 5A). Further prognostic analysis of TCGA data demonstrated that high PARP1 levels correlated with poor patient survival (Fig. 5B). Moreover, we found that plasma levels of PARP1 were significantly elevated in melanoma patients compared with healthy controls (Fig. 5C). Elevated PARP1 expression in melanoma tissues correlates with poor overall survival in patients. A, GEPIA (http://gepia.cancer-pku.cn/) was used to detect the expression of PARP1 in the TCGA melanoma dataset and the GTEx normal skin dataset. B, The prognostic data for melanoma patients in the TCGA dataset were analyzed using the GEPIA website. C, Quantification of PARP1 levels in clinical plasma specimens. *P < 0.05. TPM, transcripts per million. PARP1 Drives Functional Divergence in Melanoma Cells Using single-cell analysis of melanoma in the CancerSEA database (http://biocc.hrbmu.edu.cn/CancerSEA), we identified significantly positive correlations between elevated PARP1 expression and functional activation of cell cycle, proliferation, and DNA repair in melanoma. (See figure, Supplemental Digital Content 4, which displays how PARP1 drives functional divergence in melanoma cells, https://links.lww.com/PRSGO/F225.) Further investigation revealed cell group–specific associations: in Ma-Mel-123, PARP1 levels strongly correlated with cell cycle, proliferation, DNA damage, DNA repair, and EMT (Supplemental Digital Content 4, https://links.lww.com/PRSGO/F225); in Ma-Mel-93, PARP1 expression was primarily linked to DNA damage and repair functions (Supplemental Digital Content 4, https://links.lww.com/PRSGO/F225). These results indicated that PARP1 coordinates protumorigenic functions in melanoma with cell-specific roles. DISCUSSION Melanoma is a highly aggressive malignancy originating from melanocytes, characterized by strong metastatic potential and poor prognosis.3,4 Although recent advances in molecular targeted therapy and immunotherapy have significantly improved patient outcomes, early diagnosis, molecular stratification, and precision treatment of melanoma remain major challenges.28 Plasma proteins, as key molecular mediators reflecting systemic metabolic and immune states, play crucial roles in tumor initiation, progression, and therapeutic response. They have therefore become an important focus for understanding melanoma biology and identifying novel biomarkers.29 However, most previous studies have relied on observational data, which are susceptible to confounding and reverse causation and are often restricted by limited sample size and reproducibility. This underscores the need for genetically informed causal inference approaches to strengthen biological interpretation. MR leverages the natural random allocation of genetic variants as IVs to infer causal relationships between exposures and disease outcomes at the population level, effectively minimizing confounding and reverse causality.12 Compared with traditional proteomic association studies, MR provides greater inferential power and robustness, making it particularly suitable for identifying plasma proteins with pathogenic or protective roles. By integrating the UKB-PPP and deCODE genetics proteomic GWAS datasets, this study systematically assessed the causal effects of thousands of plasma proteins on melanoma risk. The inclusion of multilayer validation approaches, including colocalization, SMR, HEIDI, and MR-PheWAS, further enhanced the reliability of the results. This genetically anchored analytical framework offers new insights into the molecular pathogenesis of melanoma and provides a foundation for the discovery of druggable targets and the advancement of precision therapeutics. In this study, MR was first applied to infer causal relationships between plasma proteins and melanoma. To enhance result robustness and generalizability, pQTLs data from 2 independent public databases—UKB-PPP and deCODE genetics—were integrated. This analysis identified 17 plasma proteins exhibiting causal links to melanoma. Subsequent colocalization analysis was performed to determine whether the identified pQTLs and melanoma susceptibility traits share identical causal genetic variants. This validation step robustly confirmed PARP1 as the sole plasma protein causally associated with melanoma risk. SMR and HEIDI analyses using pQTL data further corroborated the role of PARP1 protein as a potential druggable target for melanoma. Concurrently, MR-PheWAS revealed minimal safety concerns, with adverse reaction risks and pleiotropic effects below critical thresholds. PARP1 is a nuclear enzyme essential for DNA single-strand break repair and genomic stability, and it also regulates chromatin remodeling, transcription, apoptosis, and inflammatory responses.13,14 PARP1 modulates tumor microenvironment dynamics through pathways, including nuclear factor kappa B.14 Furthermore, as a therapeutic target, PARP1 inhibitors leverage a “synthetic lethality” mechanism to selectively target tumor cells with DNA repair deficiencies and may also synergize with immunotherapy to improve treatment outcomes.16 Notably, elevated PARP1 expression has been shown to promote melanoma cell survival, synergize with BRAF mutations to accelerate tumor progression, and alter immune responses, thereby increasing susceptibility to melanoma.15 In addition, high PARP1 levels have also been associated with poorer overall survival in melanoma patients.30,31 Here, integrated transcriptome profiling of melanoma tissues and single-cell functional analysis of tumor cells demonstrate that elevated PARP1 expression correlates with poor clinical outcomes, specifically shorter overall patient survival. Crucially, PARP1 cell autonomously drives melanoma progression by orchestrating cell-cycle regulation, proliferation, DNA repair, and other protumorigenic processes within tumor cells. Collectively, these findings highlight PARP1 as both a potential molecular driver of melanoma progression and a therapeutic target. Currently approved molecularly targeted therapies for melanoma include BRAF inhibitors (eg, vemurafenib) and mitogen-activated protein kinase/kinase inhibitors (eg, trametinib) for patients with BRAF V600 mutations.32,33 At present, some scholars have applied PARP inhibitors to treat melanoma, but the evidence for their efficacy in melanoma remains insufficient. However, they are being actively studied as monotherapy or combination therapy in advanced/metastatic patients to explore their therapeutic potential.19 They act on the DNA damage repair pathway, distinct from the MAPK pathway targeted by BRAF/mitogen-activated protein kinase/kinase inhibitors.34 The clinical success of PARP inhibitors in other cancers, such as ovarian cancer, provides a proof-of-concept for their potential translation to melanoma.35 To integrate PARP1-targeted therapy into clinical management protocols, it is first necessary to identify whether patients harbor specific biomarkers, such as PARP1 overexpression, through tumor genomic profiling. For patients who lack effective standard targets but exhibit significant PARP1 abnormalities, PARP1 inhibitors could then be considered as an option for clinical trials or future personalized treatment. By integrating MR analysis of plasma proteomics, transcriptome sequencing of tumor tissue, and single-cell functional sequencing of tumor cells, we ensure the robust identification of biomarkers and therapeutic targets for melanoma. Owing to the high aggressiveness of melanoma, these findings demonstrate significant potential clinical relevance and public health preventive value. However, there are still some limitations. First, potential bias may arise from the relatively small sample size or overlap between plasma proteins and melanoma traits in GWAS data. Second, racial diversity was limited by the predominant focus on European populations; thus, validation across broader racial groups is necessary to confirm generalizability. Finally, the experimental validation using in vitro and in vivo models, which will include stratification by stage and subtype using our own collected samples, constitutes a critical next step toward establishing clinical translation potential. In summary, our integrative approach—which strategically combines MR, transcriptome sequencing, and single-cell functional sequencing—establishes that the plasma protein PARP1 acts as a causal risk factor driving melanoma pathogenesis. This evidence further positions PARP1 as a potential disease-specific molecular target for melanoma patients, thereby offering a viable precision medicine strategy applicable to clinical management protocols. DISCLOSURE The authors have no financial interest to declare in relation to the content of this article. Supplementary Material