针对DNA序列编码区的识别问题,本研究提出一个特征向量和逻辑回归的组合模型.首先对DNA序列进行数值处理转化为特征向量,并结合k字符相对频率技术提取特征向量的元素特征,之后利用二分类逻辑回归算法,对编码区和非编码区进行准确区分.选取了HMR195 和BG570 两个基准数据集进行五折交叉验证,结果表明,平均AUC(Area Under Curve)值分别为 0.981 3 和 0.987 4,明显优于传统的贝叶斯判别法和VOSSDFT等方法.此外,本文提出的特征向量的维度很低,提高了运算效率.因此,本文组合模型能够较为高效准确地识别蛋白质编码区.
酶分子的生物学功能很大程度上是由其三维空间结构和所处溶剂环境共同决定的.因此,优化酶分子的结构性质以及探索其性质最优的溶剂环境是改善酶分子功能以及进行理性设计的一个可行途径.从实际应用的角度来看,分子设计方法可以为酶工程提供一种有效的解决方案.目前,酶分子设计有两个重要的研究方向,包括提高酶分子的催化活力和优化其稳定性.同时,对酶分子设计方法的研究也有助于对蛋白质生物学机理的探索.在近些年的学术界酶分子设计案例中,生物信息学方法得到广泛的应用.本文系统地总结基于生物信息学的酶分子设计方法的背景、策略和一些经典案例.
以内质网应激相关基因构建骨肉瘤患者的风险模型,探索其与肿瘤免疫微环境的关系.采用生物信息学分析法,训练集的转录组数据及临床数据下载于UCSC Xena数据库,验证集的相应数据下载于GEO数据库(GSE21257,GSE39058).采用单因素COX回归分析、LASSO回归分析及多因素COX回归分析提取风险特征基因构建风险模型,使用决策曲线分析、受试者工作特征曲线分析验证模型的准确性,随后构建列线图进一步预测骨肉瘤患者预后;根据风险评分将患者分为高、低风险组,使用Kaplan-Meier生存曲线评估高、低风险组间的生存差异,对差异表达基因(Differentially expressed genes,DEGs)进行GO/KEGG联合富集分析、基因集富集分析(Gene set enrichment analysis,GSEA)及基因集变异分析(Gene set variation analysis,GSVA);采用ESTIMATE算法、微环境种群计数器(Microenvironment cell population counter,MCP counter)方法、单样本基因集富集分析(Single sample gene set enrichment analysis,ssGSEA)进行免疫分析;最终在验证集中验证上述结果.6 个风险特征基因中VEGFA、PTGIS及SERPINH1与骨肉瘤患者的不良预后相关,而TMED10、MAPK10 及TOR1B与与骨肉瘤患者的良好预后相关,高、低风险组患者之间具有显著生存差异;GO/KEGG联合富集分析、GSVA、GSEA结果表明DEGs与免疫状态相关;免疫分析显示高风险组具有更低的免疫评分及免疫景观;列线图进一步准确地预测了骨肉瘤患者的预后.内质网应激相关基因构建的风险模型能准确预测骨肉瘤患者预后,并与肿瘤免疫微环境相关.
通过生物信息学方法分析IFN-γ(Interferon-gamma,干扰素-γ)诱导银屑病皮损的关键基因及可能的作用机制.从GEO(Gene Expression Omnibus)数据库的GPL571 平台下载GSE32407 mRNA基因芯片数据集进行基因转录谱分析.设定阈值为|Log2(FC)|(差异表达倍数 2 倍的绝对值)≥1 且P＜0.05,筛选出差异基因.绘制火山图、韦恩图、蛋白质互作网络图、GO(Gene Ontology,基因本体论)/KEGG(Kyoto Encyclopedia of Genes and Genomes,京都基因和基因组百科全书)富集分析图.健康人组和银屑病病人组共筛选出1 321个DEGs(Differentially expressed genes,差异表达基因),PPI(Protein-Protein Interac-tion,蛋白质互相作用)网络筛选出ISG15、IFIT1、RSAD2、MX1、IFIT3、IFIT2、IRF7、STAT2、MX2、OASL等十个关键作用基因,国内外已有研究对IFIT3、IFIT2、OASL等 3 个基因与银屑病的关系关注较少,这 3 个基因可能成为导致银屑病的重要基因,但尚需实验验证.基于本文生信分析的预测结果,推导出IFN-γ可能通过关键基因的表达,促进角质形成细胞增殖、树突状细胞成熟和中性粒细胞浸润,导致局部炎症反应,从而导致银屑病,可为治疗银屑病的靶向药物研究和IFN-γ诱导银屑病动物模型提供一定的理论依据,但这个推论仅是通过生信分析推导的,因此还需要进一步的实验验证.
疾病关键基因可用于疾病诊断、预测和新药或新疗法有效性的评价,故识别与疾病紧密相关的关键基因十分重要.然而现在有些疾病样本数据较少,传统基于大样本的关键基因挖掘方法不适用于该类数据.本文针对含少量样本数据的疾病,首先利用单样本网络构建方法构建每个疾病样本的个体化基因网络,并通过建立基因间的层间联系构建多层基因网络.然后利用基于张量的多层网络中心性方法评估每层网络中基因间的相互作用以及层间影响,对基因进行重要性打分,识别疾病关键基因.最后将该方法应用到哮喘数据集上,并与经典算法进行比较,结果表明,利用该方法所识别的已获批准的药物靶标基因的排名较优;对所得到的新的潜在关键基因TP53、PUS10、MAP3K1 等进行功能和通路富集分析,结果表明其与哮喘有紧密关联.
应用网络药理学及分子对接技术探究槲皮素治疗抑郁的潜在机制.从SwissTargetPrediction、Superpred数据库中筛选出"槲皮素"相关靶点;从OMIM、Genecards数据库中筛选出"抑郁"的相关靶点.通过Jvenn获得二者的共同蛋白质靶点信息后,使用Cytoscape软件构建蛋白质互作网络,并对hub基因进行筛选.使用David平台对相同靶点进行GO、KEGG等富集分析.最后采用分子对接技术进行准确性检验.PPI网络中共有 103 个节点,536 条边,其中SRC、MTOR、EGFR、AKT1、PTK2 等靶点度值排名较高.GO、KEGG富集分析结果表明,槲皮素对抑郁的作用主要涉及凋亡过程的负向调节、信号转导、蛋白磷酸化反应等生物学过程.信号通路主要包括HIF-1、ErbB、磷脂酶D信号传导、神经营养素信号传导等.分子对接结果显示SRC、MTOR、AKT1 等靶点与槲皮素结合程度较好,且在KEGG通路中富集.揭示了槲皮素作用于抑郁的潜在靶点及机制,以期望研制出治疗抑郁症新药物.
通过比较登革热患者和健康人群转录组数据,识别差异基因,构建失调ceRNA网络,筛选关键基因富集分析,解析潜在生物学功能,助力登革热诊断标志物的研究.从GEO数据库下载登革热外周血芯片数据,识别差异基因并进行富集分析.结合miRNA-mRNA互作数据,利用超几何算法和皮尔森相关性计算方法识别登革热失调ceRNA互作对,使用Cytoscape软件可视化ceRNA网络与模块挖掘,对网络模块进行功能富集及外部数据验证表达模式.筛选出 251 个差异基因,发现其富集在细胞周期等生物学通路中.经外部数据验证,网络模块基因的表达趋势与训练集数据大致相同,表明模块基因在登革热疾病中的潜在诊断效能.本研究可为确定有效的疾病诊断分子标志物提供思路.
为寻找与家族性双侧大结节性肾上腺皮质增生症发展有关的潜在治疗靶点和生物标志物.从GEO数据库中下载GSE171558 数据集,筛选受家族影响的肾上腺结节与正常的肾上腺组织之间的差异表达基因(Differentially expressed genes,DEGs),并进行基因功能富集分析和蛋白质-蛋白质相互作用网络分析.通过Cytoscape v3.9.1 软件中的插件cytoHubba筛选出关键基因,进一步经NetworkAnalyst分析TF-miRNA共调控网络和蛋白质-化合物相互作用.共鉴定出336 个DEGs,这些基因主要富集在细胞粘附过程、细胞增殖的正调节过程和RNA加工过程等生物过程,并涉及钙信号通路、PI3K-Akt信号通路和cAMP信号通路等.通过cytoHubba插件获得 5 个hub基因,经验证分析,多功能蛋白聚糖(Versican,VCAN)、双糖链蛋白聚糖(Biglycan,BGN)被认为是家族性双侧大结节性肾上腺皮质增生症的潜在生物标志物.进一步的GSEA分析结果显示,VCAN主要与丁酸代谢、ECM-受体相互作用和类固醇生物合成等有关.BGN主要涉及剪接体、皮质醇的合成和分泌和类固醇生物合成等.利用NetworkAnalyst分别展示了生物标志物与miRNA和已知化合物的相互作用网络.VCAN和BGN基因可作为家族性双侧大结节性肾上腺皮质增生症的潜在生物标志物,这些基因特征和通路可能是家族性双侧大结节性肾上腺皮质增生症患者新的治疗靶点.
本文提出了一种新的快速非比对的蛋白质序列相似性与进化分析方法.在刻画蛋白质序列特征时,首先将氨基酸的10 种理化性质通过主成分分析浓缩为 6 个主成分,并且将每条蛋白质序列里的氨基酸数目作为权重对主成分得分值进行加权平均,然后再融合氨基酸的位置信息构成一个26 维的蛋白质序列特征向量,最后利用欧式距离度量蛋白质序列间的相似性及进化关系.通过对3 个蛋白质序列数据集的测试表明,本文提出的方法能将每条蛋白质序列准确聚类,并且简便快捷,说明了该方法的有效性.
Equus przewalskii belongs to the national first-class protected wild animals. It is the only surviving wild subspecies of horse. Growth hormone（GH） is a polypeptide hormone secreted by the anterior pituitary gland, which has an obvious effect on the growth and development of animals. In order to better protect the wild horse and improve its wild survival and breed ability, the GH gene of Equus przewalskii was obtained by PCR for the first time, and the structure and function of its encoded protein were analyzed and predicted by bioinformatics method, so as to explore the structural characteristics of GH gene and the structure and function of its encoded protein. Results showed that the total length of genomic DNA of GH gene of Equus przewalskii was 1577 bp. After BLAST comparison, it was found that the sequence similarity of GH gene between Equus Przewalskii and Equus caballus（GenBank: EU939446.1） was as high as 99.68%, indicating that they belong to homologous sequence. The number of amino acids of the protein encoded by the gene was 152, the molecular weight was 17 638.54 Da, the theoretical isoelectric point was 9.01, the total number of negatively charged residues（Asp+Glu） was 19, the total number of positively charged residues（Arg+Lys） was 23, and the molecular formula was C 79 0H 1274 N 216 O 226 S 7 . GH protein of Equus przewalskii belonged to extracellular protein, which contained growth hormone like domain and did not have relevant transmembrane structure, with a 0.35% chance of containing signal peptide. The research lays a molecular foundation for the development and utilization of GH gene in Equus przewalskii.
Machine learning was used for idenfifying possible pathogenesis-related long non-coding RNAs（lncRNAs） in labial gland tissue and whole blood of Sjogren’s syndrome（SS） patients, and investigating the correlation between the lncRNA with immune cell infiltration in labial gland tissue. The expression data of SS labial gland tissue and whole blood were obtained from the GEO database, and the data were normalized and processed for differential analysis to obtain lncRNA expression profiles. Common lncRNAs were screened and intersected by two machine learning methods. Based on CIBER-SORT software, 22 immune cell infiltration situations in labial gland tissues were calculated, and the correlation between key lncRNAs and immune cell infiltration was analyzed. The correlation between key lncRNAs and clinical traits in whole blood was analyzed and ROC curves were plotted. Key lncRNA co-expression encoding protein genes were analyzed and KEGG signaling pathway analysis was performed. One SS key lncRNA（HCP5） was identified, and HCP5 was upregulated in labial gland tissue and whole blood. Immuno-infiltration analysis showed that the proportion of γδ T cells, macrophages, and CD4 + memory T cells increased in SS labial gland tissue, while the proportion of plasma cells decreased in SS labial gland tissue. The expression of HCP5 was positively correlated with the infiltration of γ/δ T cells and CD4 + memory T cells. In whole blood, HCP5 significantly correlated with ANA, IgG, anti-SSA antibodies, anti-SSB antibodies, and ESSDAI. The ROC diagnostic curve results in different datasets showed that the AUCs of HCP5 were 0.833 and 0.877. KEGG analysis showed that the function of proteins co-expressed with HCP5 in the labial gland was focused on signaling pathways such as antigen processing and presentation and B cell receptor signaling pathway, while the proteins co-expressed with HCP5 in whole blood functioned in signaling pathways such as metabolic pathways and apoptosis. HCP5 may affect the pathogenesis and progression of the disease by regulating the infiltration of immune cells in the labial gland tissue, and may be a potential diagnostic and therapeutic target for SS.
Base editors are practical and efficient gene editing tools, whose editing efficiencies often depend on the design of single guide RNA（sgRNA） sequences. At present, the design of sgRNA libraries lacks of specific rules and mainly relies on experience and attempts. On the basis of the convolutional neural network, BEguider was developed for sgRNAs design of base editors. BEguider used the deep learning framework TensorFlow 2 to build editing efficiency prediction models, which could design sgRNA sequences and predict editing probabilities for NGG PAM-dependent base editor variants ABE7.10-NGG and BE4-NGG within the scope of the human genome. Besides, Beguider could evaluate potential off-target sites of sgRNAs by integrating Cas-OFFinder. Using BEguider to design sgRNA sequences will facilitate future application of base editors and save experimental cost.
In order to predict and analyze the structure and function of human CREB-binding protein（CBP）, bioinformatics methods were used to predict the physical and chemical properties, conservation, subcellular localization, signal peptide, transmembrane domain, secondary structure, tertiary structure, interacting protein, and function of human CBP. Results showed that human CBP was an unstable and hydrophilic protein located in the nucleus without transmembrane region and signal peptide. The secondary structure was characterized by random coil and α-helix. The HAT domain of the protein was highly conserved among species. It was speculated that the amino acid residues closely related to its enzyme activity were Tyr1433, Leu1434, Asp1435, and Arg1664. In addition, human CBP could interact with transcription factors and transcription coactivators, such as TP53, CREB1, and NCAO3, and was mainly involved in biological processes such as transcription regulation, cell differentiation, tissue development, signal transduction, and apoptosis. This study provides a theoretical basis for further studying the mechanism of CBP in the occurrence and development of malignant tumors.
Delayed graft function（DGF） is one of the common complications of kidney transplanta patients. More and more studies have begun to focus on new pathophysiological mechanisms and potential diagnostic markers of DGF after kidney transplantation. In this study, the gene expression profile dataset of kidney transplant patients in the GEO database was analyzed. Through differentially expressed genes（DEGs） screening, the dysregulated expression of multiple transcription factors and immune genes was found, and the core regulatory genes in the process of disease progression were further explored through the interaction network analysis between gene-encoded proteins. The prediction model of DGF kidney renal transplantation was constructed by combining weighted gene co-expression network analysis（WGCNA） and machine learning. The accuracy of model XGBoost reached 82.4%, its area under the receiver operating characteristic curve（AUC） was 0.86, Matthews correlation coefficient（MCC） was 0.652, and sensitivity and specificity were 0.789 and 0.867 respectively. Retrieval of these characteristic genes with optimal predictive power found that these genes were closely related to renal function. Finally, several small molecular compounds that could be used to treat DGF were found by comparing the CMap database. This study explored the pathophysiological mechanism of DGF from multiple perspectives, providing a reliable theoretical and experimental basis for the diagnosis and treatment of related diseases.
The functional affiliation and inner relationships between the risk factors and the mechanism of living organisms were analyzed based on the relevant interactive diagrams drawn from the analysis of the information characteristics of the risk factors and the mechanism of living organisms. The two-dimensional relevant interactive diagrams were upgraded to a three-dimensional diagram, which constructs the life information security control by Tai Chi mapping of theodolite. The Tai Chi mapping of theodolite is composed of five Tai Chi latitude planes, and a life information security control axis that runs through the center of the five Tai Chi latitude planes. The control axis and the Tai Chi latitude planes form the relationship of abstract and real and the relationship of longitude and latitude, which forms a situation of “attain the utmost in passivity, hold firm to the basis of quietude”, that is, a state of dynamic balance. The five Tai Chi latitude planes represent pattern recognition receptors to recognize pathogen-related molecular patterns and damage-related molecular patterns, NK cell activation receptor recognition mechanism, specific recognition system mechanism, positive and negative feedback regulation mechanism, and the interaction mechanism of mucosa, skin, and symbiotic microorganisms. Intrinsic functional linkage analysis showed that the first and second Tai Chi latitude plane mechanisms were at the forefront of the entire system, which had the first response to the dangergen, and the dangergen information was transmitted to the third Tai Chi latitude plane; the third Tai Chi latitude plane accepted the dangergen information and generated a second response, in the “backup position”; the fourth Tai Chi latitude plane made timely adjustments for excessive or insufficient information response, and performed positive and negative feedback functions, which played an extremely important role in the overall system; the fifth Tai Chi latitude plane mechanism was the foundation of the entire safety control system, especially the mucosal system. The key layout of the safety control system was concentrated in the fifth plane. The large amount of symbiotic microorganisms embodied in the layout of functional forces determines its important role. The Tai Chi mapping of theodolite showed that the ligands in the NK cell activation receptor（receptor for answeringstress-response information）-ligand（stress-response information） relationships are included in the system.The concept of “dangergen” was firstly proposed, which covers traditional antigens, pathogen-related molecular patterns, damage-related molecular patterns and the ligands in the NK cell activation receptor-ligand.The hypothesis of the principle of NK cell function effect mechanism was also proposed: on the basis of the overall genetic information of the species, the relationship between the supporting receptors and ligands that determines whether to attack or not was constructed, and it was speculated that there were gene families in the cell genome that express ligands representing their own identity information. and those express reporter allergenic information ligands.The constructed Tai Chi mapping of theodolite shows the overall conformation of the life information security control mechanism.
On the basis of the acute myeloid leukemia（AML） clinical data and multi-omics database, the role of ferroptosis-related genes in AML was explored, and a prognostic model related to gene expression of ferroptosis was established. The clinical and transcriptomic datasets of AML and controls were obtained from TCGA [AML（n=151）] and GTEx [whole blood（n=337）]. The Wilcoxon test and univariate Cox analysis results were intersected to screen out differentially expressed genes（DEGs） related to prognosis, and Lasso regression was used to build a gene signature prognostic model. Receiver operating characteristic（ROC） curve was used to evaluate the predictive value of the model, the Kaplan-Meier method was utilized for survival analyses, univariate and multivariate Cox regression analysis of clinical data were performed, and differential gene expression analysis and other methods were adopted to compare transcriptomic differences between high and low risk patients. Finally, the gene signatures were validated using the BeatAML database. Results of differential gene expression analysis and univariate analysis were intersected to obtain 13 prognosis-related DEGs. A prognostic scoring model of eight genes was constructed, and the patients were divided into high and low risk groups. ROC curve analysis confirmed the good predictive performance of the model, survival analysis showed that the survival rates of patients in the high and low risk groups were significantly different, univariate analysis revealed that age and risk score were significantly associated with overall patient survival, and multivariate analysis demonstrated that age and risk score were independent prognostic indicators. A total of 384 DEGs were screened between the two risk groups. The GO enrichment analysis showed that most of the enriched genes were related to immune-related molecules and pathways such as chemotaxis and migration of neutrophils and leukocytes. The KEGG enrichment pathway was mainly related to the TNF signaling pathway, cytokine and cytokine receptor interaction. The validation using the BeatAML database showed that five genes were significantly associated with prognosis. Ferroptosis-related genes are significantly expressed in AML, and high risk patients have a poor prognosis. This study lays a foundation for the discovery and application of potential biomarkers related to ferroptosis in AML.
Lipid metabolism is an important metabolic process of the body, and its disorder will lead to many diseases. Cluster of differentiation 36（CD36） is a scavenger receptor highly expressed in monocytes, macrophages, smooth muscle cells and adipocytes. It is the main receptor and transporter for the recognition of oxidized low-density lipoprotein or long-chain fatty acids. So it plays an important role in the process of lipid metabolism. The review recapitulates the update and current advances on the structure and function of CD36 gene and protein, expounds the role of CD36 in the process of lipid metabolism, and systematically summarizes the molecular mechanism that CD36 cascade AMPK, mTOR and MAPK signaling pathways to participate in the process of lipid metabolism, providing a theoretical basis for the relevant biological research.
The mechanism of an autophagy-related gene（ARG） in the occurrence and development of multiple myeloma（MM） was explored and a related prognostic model was established. On the basis of the MMRF and HADb databases, the differential expression of ARGs in MM was determined by R language, the relationship between ARGs and the occurrence and development of MM was analyzed by GO and KEGG, the multi-gene prognostic model was established by COX regression algorithm, the survival curve was drawn by Kaplan-Meier method, and the reliability of the prognostic model was evaluated by ROC curve. Finally, a total of 104 differentially expressed ARGs were extracted from bone marrow samples of 764 MM patients and four normal bone marrow samples. It was found that there were significant differences in the expression of 104 genes in MM samples, including 46 up-regulated genes and 58 down-regulated genes. GO enrichment mainly focused on ontology annotations such as macrophagy, autophagy regulation, and cell response to external stimuli. KEGG enrichment was mainly concentrated in autophagy, apoptosis, NOD-like receptor signal pathway, and PI3K-Akt signal pathway. Univariate COX analysis showed that 33 ARGs were significantly correlated with the overall survival of MM patients. Thirteen prognostic ARGs（NKX2-3, NCKAP1, BIRC5, PEX3, HGS, RUBCN, PARP1, ARSA, DNAJB9, HSP90AB1, EEF2, FKBP1B, and CD46） were selected by multivariate COX regression to establish ARGs prognostic model for MM. Kaplan-Meier survival curve analysis showed a significant difference in survival rate between high risk group and low risk group（P ＜0.001）. Multivariate COX regression analysis showed that age, ISS stage, and risk value were independent prognostic indicators of MM patients（P＜0.005）. The area under the ROC curve was 0.561, 0.685, and 0.719. The following conclusions are drawn: the ARGs prognostic model of MM can be used to predict the survival of patients with MM, but it needs to be verified by further clinical studies.
探讨铁死亡相关基因在肾透明细胞癌患者中的表达及其预后价值.通过TCGA数据库下载KIRC的相关测序数据与检索到的铁死亡相关基因取交集,进行铁死亡相关基因的差异分析.之后利用单变量和多变量Cox回归分析,筛选具有预后价值的基因,构建预测患者生存情况的风险评分模型,并对模型进行验证.对高低风险组进行GO与KEGG通路富集,探讨风险差异的可能原因;通过ssGSEA分析,评估高低风险组间的免疫浸润情况.在KIRC患者的肿瘤组织和正常组织中,共得到21个差异的铁死亡相关基因;通过单因素Cox回归分析,获得28个与KIRC预后相关的基因;之后进行Lasso回归与多因素Cox回归分析,结果显示有10个基因被纳入模型,计算公式为:风险值(Risk score)=(0.0245)×ALOX5表达值+(0.1260)×CBS表达值+(0.1995)×CD44表达值+(0.2183)×CHAC1表达值+(-0.2959)×HMGCR表达值+(0.0367)×MT1G表达值+(0.0614)×SLC7A11表达值+(-0.0807)×FDFT1表达值+(0.1603)×PEBP1表达值+(-0.2205)×GOT1表达值.生存状态图表明,高风险组死亡病例数多于低风险组;ROC曲线表明风险评分模型具备一定预测能力;K-M生存分析显示,高风险组总体生存率低于低风险组(P=5.73×10-13).GO与KEGG富集分析提示,高低风险组间免疫情况及IL-17信号通路存在显著差异;进一步的ssGSEA富集显示,高低风险组间大部分免疫细胞的评分存在显著差异.基于铁死亡相关基因的预后风险评分模型可用于KIRC的预后预测,针对铁死亡相关基因设计靶点可能是治疗KIRC的一种新选择.
结合TCGA数据库中宫颈癌的lncRNA表达谱和体细胞突变谱,构建基于突变假设的计算框架,鉴定出36个与宫颈癌基因组不稳定性相关的lncRNA;对其共表达的基因功能进行分析,发现与36个lncRNA共表达的基因在2-氧代戊二酸代谢过程和2-氧羧酸代谢通路中富集.构建了基于基因组不稳定性衍生的两个lncRNA的基因特征(GILncSig),将Train组患者分为高风险组和低风险组,两组患者生存率显著不同,这一结果在Test组患者中得到进一步验证.通过独立预后分析,结果显示GILncSig可独立于其他临床性状,作为宫颈癌患者的整体生存相关独立预后因子.总之,本研究为进一步探讨lncRNA在基因组不稳定性中的作用提供了关键的方法和资源,为识别基因组不稳定性相关的肿瘤标志物提供了新的预测方法.
