基于逐步模型树多组学融合调控网络方法的直肠癌新辅助化疗敏感标志物筛选研究

术前新辅助治疗作为直肠癌患者的重要治疗手段，可以让不可切除的肿瘤变得可切除，提高患者保肛率，改善患者生活质量，优化无瘤生存率和总生存率。但并不是所有接受新辅助治疗的直肠癌患者都能获得良好的病理完全反应继而从中获益，有研究显示完全病理反应的患者仅占10%-25%。因此如何识别对新辅助治疗反应敏感患者，避免不敏感患者的过度治疗和无效治疗，是目前临床难题，也是个体化医疗的面临的重大挑战。.本研究提出逐步模型树多组学融合调控网络模型，对多组学数据建立有方向且加权的生物调控网络模型。进而利用此模型融合分析新辅助化疗直肠癌患者的蛋白质组学和代谢组学数据并建立调控网络，以发掘更具有临床意义和更接近生物学实际的化疗敏感标志物和相关代谢通路。为临床直肠癌新辅助化疗敏感患者的预测和识别提供强有力证据，为直肠癌患者选择合适的治疗方案，进一步加强个体化医疗提供有力支持。

Neoadjuvant chemotherapy (NCT) is the standard therapy for rectal cancer patients (NCPs), letting it possible to resect rectal tumors at terminal stages, enhancing the probability of anus reservation, improving the life quality, tumor-free survival and overall survival. However, after accepting the NCT, not all of NCPs would be sensitive and beneficial from it, and even members of pathological complete remission only account for 10%-25%. Indeed, it is also a difficult problem for clinical therapy and individualized medicine practice that how to distinguish heterogeneous NCPs before treatment, and avoid overtreatment and ineffective treatment for the non-sensitive. .In our study, we make an analytical model that takes directions and weights in regulatory networks into account, that is multi-omics fusion regulatory network based on progressive model tree. Further, we could utilize it to fuse proteomics and metabolomics data and build their regulatory networks, and then discover NCT-sensitive biomarkers and relative metabolic pathways. Through method of data fusion, we can select biomarkers and pathways with more clinical and biological meanings, and provide us more evidences to distinguish NCT-heterogeneous patients, in order to choose proper therapeutic regimen and improve effectiveness of individualized medicine.