基于AFM的单细胞多维信息分析

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 机器人学研究室

题名	基于AFM的单细胞多维信息分析
作者	周富元
学位类别	硕士
答辩日期	2016-05-25
授予单位	中国科学院沈阳自动化研究所
导师	王文学
关键词	原子力显微镜细胞机械特性自动化
其他题名	Analyzing the multiple information of single cell obtained by AFM
学位专业	机械电子工程
中文摘要	单细胞包含了生命体丰富的信息。2012年哈佛大学的谢晓亮发表在《Science》杂志中的文章指出：利用单个细胞获取了整个生命体的遗传信息。单细胞生物学信息的研究对药物筛选，疾病的诊断、个性化医疗都具有重要的意义。在单细胞单分子水平上的多维生物学信息的研究，正成为生物医学领域的新方法。在众多科学家的共同努力下，细胞生物学信息的研究成果已经硕果累累，但是还有一些问题需要解决。其一，现在治疗疾病的药物存在着疗效差异，严重阻碍着个性化医疗的发展。以淋巴瘤为例，美罗华是一种治疗淋巴瘤的靶向药物。美罗华的使用可以缓解大部分患者病情改善他们的生活质量，但是仍约有25%的患者不但不能获得病情的缓解，反而出现内分泌失调、恶心等副作用。为了研究其耐药性的原因，就需要对美罗华分子的作用靶点淋巴瘤细胞表面的CD20分子进行探测。而利用原子力显微镜对CD20分子进行探测的过程中会得到三类分子结合力曲线，其中只有特异性分子结合力曲线是由于CD20分子与美罗的结合产生的。因此需要从这三类力曲线中辨识出特异性分子结合力曲线，目前分子结合力曲线的识别主要靠人工，效率低，因此有必要实现分子结合力曲线的辨识自动化。其二，目前细胞杨氏模量的计算主要采用基于Hertz 模型的高斯拟合法，该方法无法保证得到的理论压痕曲线和实验压痕曲线拟合得很好，这就无法保证得到的杨氏模量值的正确性。其三，目前细胞机械特性的研究大多集中在细胞杨氏模量上，缺乏从多维细胞机械特性角度综合分析细胞状态。本论文研究以国家自然科学基金项目(61327014,61433017)为依托，针对以上存在的几个问题进行了研究：针对单分子力曲线自动识别问题，对在实验中得到3类分子结合力曲线，结合单分子力曲线的特点，利用数据挖掘方法，实现单分子力曲线的自动识别。针对基于Hertz模型的杨氏模量计算的高斯拟合法无法使理论压痕曲线与实验压痕曲线重合问题，提出一种最优化方法，该方法能使理论压痕曲线与实验压痕曲线拟合最好，并以MCF-7细胞为实验对像，验证方法的有效性。针对细胞状态表示的问题，证明多维机械特性相比于单一的细胞机械特性表示细胞状态的优越性，并以MCF-7 细胞为实验对象，利用多维细胞机械特性（杨氏模量、能量耗散）表示细胞状态，研究了不同药物浓度下MCF-7的状态的变化过程。以上研究，将为细胞机械特性的计算，为药物的筛选和个性化的医疗奠定基础。
英文摘要	Cells contains rich information about state of life body. In 2012, the scientist Xie obtained the whole genetic information about body from single cell, which published in science magazine. The study of information of single cell will be useful for drug selection, disease diagnosis, and personal treatment. The study of information of single cell will provide methods for biomedical science. With the efforts of scientists, there are many important findings about the information of single cell. But there are still many problems remained to be solved. Firstly, there are many drugs which efficacy is different among patients which brings obstacles to the development of personalized treatment. Take the lymphoma for example. Rituximab has targeted therapy of lymphoma. The rituximab can improve the state of lymphoma and the quality of life of patients. But there are a proportion of patients has side effects. In order to figure out the reasons, there is a need to detect the CD20 molecule. There are three types of force curves when we use the AFM to detect the CD20 molecule. Of the three types of force curves only the special binding force curves formed because of CD20 binding with rituximab. So in order to detect the CD20 molecule, there is a need to recognize the special binding force curves. However manual recognition of the force curves is time-consuming and is not realistic especially for large dataset. Therefore there is a need to realize automatic identification of the force curves. Secondly, the Gaussian fitting method cannot make sure that the theoretical indentation curve fit the experimental indentation curve well, which means that we cannot make sure the correctness of the computed young’s modulus. Thirdly, many researchers often only considered the young’s modulus and seldom considered multiple cell mechanics to represent the cells’ state. This study was supported by the National Natural Science Foundation of China (Grant No. 61327014, Grant No. 61433017). We studied the several problems described above in this thesis. For the needs of automatic classification of force curves. We demonstrated a data mining based approach to automatic classification of the three types of force curves. For the problem that the Gaussian fitting method cannot make sure that the theoretical indentation curve fits the experimental indentation curve well. We proposed an optimization method which can make the theoretical indentation curve fit the experimental indentation best. For proving correctness of the optimization method, an experiment was carried out. We use the MCF-7 cells for the experiment. For the problem of representing the state of cells, we demonstrated that using multiple cell mechanics to represent the state of cells is better than using only one of the cell mechanics. To prove this, we use the multiple cell mechanics to represent the state of MCF-7 cells and the MCF-7 cells was used for the experiment which has been cultured by different concentration of resveratrol for 36 hours. The research described above will lay a foundation for the computation of cell mechanics, drug selection, disease diagnosis and personalized treatments.
语种	中文
产权排序	1
页码	61页
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/19671]
专题	沈阳自动化研究所_机器人学研究室
推荐引用方式 GB/T 7714	周富元. 基于AFM的单细胞多维信息分析[D]. 中国科学院沈阳自动化研究所. 2016.