第一部分：

背景：肺腺癌（lung adenocarcinoma，LUAD）是非小细胞肺癌（non-small cell lung cancer，NSCLC）中最常见的亚型，通常伴随代谢异常。嘌呤生物合成受磷酸核糖甲酰胺咪唑合成酶（phosphoribosylaminoimidazole carboxylase and succinocarboxamide synthetase，PAICS）调控，在肿瘤进展和治疗耐药中起重要作用。

方法：本研究利用泛癌分析和京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes，KEGG）富集分析聚焦于肺腺癌。通过分析癌症基因组图谱（The Cancer Genome Atlas，TCGA）-LUAD及三个基因表达综合数据库（Gene Expression Omnibus，GEO）数据集，确认嘌呤生物合成的预后相关性。利用最小绝对收缩和选择算子（Least Absolute Shrinkage and Selection Operator，LASSO）回归构建并在独立队列中验证了一个预后模型，即嘌呤生物合成相关评分（Purine Biosynthesis-Related Score，PBRS）。此外，研究还开展了基因集变异分析（gene set variation analysis，GSVA）、免疫特征分析、肿瘤突变负荷（tumor mutational burden，TMB）分析以及药物敏感性分析。进一步在LUAD组织中验证PAICS表达水平，并通过细胞增殖和迁移实验评估其作用。

结果：PBRS模型将LUAD患者分为高风险（PBRS-high）和低风险（PBRS-low）两个亚组，其预后存在显著差异。高PBRS患者富集于细胞周期调控和DNA修复通路，且具有较高的TMB，提示可能对免疫治疗敏感，但同时存在免疫逃逸机制可能限制免疫检查点抑制剂（immune checkpoint inhibitors，ICIs）的疗效。低PBRS患者对代谢抑制剂治疗反应更为敏感。此外，PAICS高表达与较差的预后显著相关，而敲低PAICS可显著抑制LUAD进展。

结论：PBRS可作为肺腺癌预后的有效工具，PBRS高表达患者可能从免疫治疗或DNA损伤修复（DNA damage repair，DDR）靶向治疗中获益，而PBRS低表达患者对代谢抑制剂敏感。此外，PAICS具有作为治疗靶点的潜力。

第二部分：

背景：肉瘤是一类来源于中胚层的异质性肿瘤，且对系统治疗的反应普遍较差。泛素化是一种翻译后修饰方式，参与多种生理过程并在肿瘤发生发展中发挥重要作用。

方法：本研究整合了癌症基因组图谱（The Cancer Genome Atlas, TCGA）中256例肉瘤患者的转录组数据，对预定义的泛素化相关基因的表达情况展开分析，并据此识别出两个泛素化相关亚群。随后，进行生存分析、富集分析以及肿瘤微环境分析。进一步构建了一个包含LRRC41、RNF125、TRIM21和UBE3D的泛素化相关模型，用以预测预后。该模型能够将患者划分为高风险组和低风险组，其预后价值在公共数据集以及本中心的独立队列中均得到验证。此外，通过细胞实验探究了TRIM21在肉瘤进展过程中的作用。

结果：这两种亚群在生存结局、富集通路和肿瘤微环境特征上有所不同。该模型能够有效预测公共肉瘤队列和我们队列中患者的预后。免疫组化分析表明，风险评分和CD8的结合能够更好地区分我们队列中不同生存结局的患者。从机制上讲，不同风险组还表现出不同的富集通路和肿瘤微环境特征。细胞实验揭示，TRIM21过表达可以抑制肉瘤的肿瘤进展。

结论：本研究提供了基于泛素化相关基因表达的新型肉瘤患者分类方法。该分类和预后模型有助于理解肉瘤的发病机制、预测预后和免疫治疗反应。同时，我们确认TRIM21抑制肉瘤进展，并将其确定为潜在的治疗干预靶点。

PART Ⅰ ：

Background: Lung adenocarcinoma is the most common NSCLC and is associated with metabolic dysregulation. Purine biosynthesis, regulated by PAICS, plays a key role in tumor progression and therapy resistance.

Methods: We focused on LUAD using pan-cancer and KEGG enrichment analyses. TCGA-LUAD and three GEO datasets were analyzed to confirm the prognostic relevance of purine biosynthesis. A prognostic model, the Purine Biosynthesis-Related Score (PBRS), was developed using LASSO regression and validated in independent cohorts. Gene set variation analysis, immune profiling, tumor mutational burden analysis, and drug sensitivity analysis were conducted. PAICS expression was validated in LUAD tissues, and its role was assessed via proliferation and migration assays.

Results: PBRS classified LUAD patients into high-risk (PBRS-high) and low-risk (PBRS-low) subgroups, with distinct prognostic outcomes. PBRS-high patients showed enrichment in cell cycle regulation and DNA repair pathways and had higher TMB, suggesting potential sensitivity to immunotherapy, although immune escape mechanisms may limit the efficacy of immune checkpoint inhibitors. PBRS-low patients were more responsive to metabolic inhibitors. PAICS overexpression correlated with poor prognosis, while its knockdown suppressed LUAD progression.

Conclusion: PBRS is a prognostic tool in LUAD, identifying PBRS-high patients who may benefit from immunotherapy or DDR-targeted therapies. PBRS-low patients exhibit sensitivity to metabolic inhibitors. PAICS is a potential therapeutic target.

PART Ⅱ：

Background: Sarcomas are heterogeneous tumors deriving from the mesenchyme and have a poor response to systemic therapies. Ubiquitination is a post-translational modification that is involved in various physiological processes and cancer growth.

Methods: We integrated the transcriptome data of 256 sarcoma patients from The Cancer Genome Atlas (TCGA) to analyze the expression of the predefined ubiquitination-related genes and identified two ubiquitination-related clusters based on that. Survival analyses, enrichment analyses, and tumor microenvironment analyses were conducted. Further, a ubiquitination-related model involving LRRC41, RNF125, TRIM21, and UBE3D was built to predict prognoses. The model could divide patients into high or low risk groups and the prognostic value of the model was validated in public datasets and an independent cohort from our center. In addition, the role of TRIM21 in sarcoma progression was explored using cellular experiments.

Results: The two clusters were characterized by different survival outcomes, enriched pathways, and characteristics of the tumor microenvironment. The model could effectively predict patient prognoses across the public sarcoma cohort, and our cohort. Immunohistochemistry analyses revealed that the combination of risk score and CD8 could better distinguish patients with different survival outcomes in our cohort. Mechanistically, different risk groups were also characterized by distinct enriched pathways and characteristics of the tumor microenvironment. Cellular experiments revealed that TRIM21 overexpression could suppress tumor progression in sarcomas.

Conclusions: This study provided novel classification for sarcoma patients based on expressions of ubiquitination-related genes. The classification and the prognostic model may facilitate the understanding of sarcoma pathogenesis, prediction of prognosis and immunotherapy response for sarcoma patients. Meanwhile, we confirmed that TRIM21 suppressed sarcoma progression and identified it as a potential target for therapeutic interventions.