摘要
乳腺癌是女性最常见的恶性肿瘤之一,其发展是一个长期的、多阶段、多基因改变积累的过程。临床及实验研究发现,在乳腺癌发展与转移的过程中,许多原癌基因、抑癌基因及相关蛋白质发生改变。本文提出了一种结合随机游走重启算法的综合模型,并将其应用于乳腺癌骨转移关键候选基因的鉴定中,在蛋白质相互作用网络上识别乳腺癌转移候选基因。首先利用RWR算法和置换检验规则对基因进行预选,得到候选基因集。然后构建一个候选基因间的关键子网并计算关键子网中节点的介数中心性。最后,利用交互得分规则对基因筛选,选择介数得分排名前三十的基因作为乳腺癌骨转移关键基因。相互作用分析、富集分析及文献挖掘的结果表明26个潜在的关键基因都参与了乳腺癌骨转移的起始或进程,验证了该综合模型的有效性。
Breast cancer is one of the most common malignant tumors in women, and its development is a long-term, multi-stage, multi-genetic alteration accumulation process. Clinical and experimental studies have revealed that many proto-oncogenes, oncogenes and related proteins are altered during the process of breast cancer development and metastasis. In this paper, we propose an in-tegrated model combining random wander restart algorithm and apply it to the identification of key candidate genes for breast cancer bone metastasis on protein interaction network to identify breast cancer metastasis candidate genes. The genes are first pre-selected using the RWR algorithm and the substitution test rule to obtain the candidate gene set. Then a key subnet among candidate genes is constructed and the betweenness centrality of nodes in the key subnet is calculated. Finally, the genes were screened using the interaction score rule, and the top thirty genes in the mediator score were selected as the key genes for breast cancer bone metastasis. The results of interaction analysis, enrichment analysis and literature mining indicated that 26 potential key genes were involved in the initiation or progression of breast cancer bone metastasis, validating the validity of this comprehensive model.
出处
《理论数学》
2023年第5期1267-1280,共14页
Pure Mathematics
