Circulating tumor cells （CTCs） present in the blood play important roles in cancer recurrence and metastasis， so the capture of CTCs has become a key issue in current cancer research. The method of using antibody-dependent immunomagnetic beads to capture CTCs has the characteristics of simple operation and rapidity， but there are problems such as large sample size requirement and high preparation cost. This study intended to use a new type of magnetic bead-plasma-modified graphene-coated iron nitride magnetic beads （G@FeN-MB） to carry out a preliminary study on the capture of circulating tumor cells in lung cancer. G@FeN-MB is covered by multi-layer graphene， and the core is iron nitride， which has stronger magnetic response than the commonly used iron oxide core. Graphene used as a magnetic bead coating shell has lower mass， higher stability and better biocompatibility than the SiO₂ shell commonly used for magnetic beads on the market. In this study， two kinds of streptavidin magnetic beads were prepared by direct adsorption method and carbodiimide method respectively， and coupled with CTCs surface markers epithelial cell adhesion molecule （EpCAM） antibody preparation amino immunomagnetic beads （AIMB） and carboxylic immunomagnetic beads （CIMB）. The results showed that both kinds of immunomagnetic beads could capture A549 cells， and the capture efficiency of AIMB was slightly higher than that of CIMB； in order to simulate the CTCs capture environment， healthy human blood was mixed with A549 cells to prepare cell mixtures of different concentrations， and the capture efficiency was higher by using AIMB captures A549 in human blood， and calculated the captured A549 and capture efficiency. The results showed that the capture efficiency of AIMB ranged from 66.25% to 81.50%. The feasibility of the method of preparing immunomagnetic beads based on G@FeN-MB was proved， and the obtained AIMB can be used for the sorting of lung cancer CTCs， showing the prospect of its application in clinical research.

In order to investigate the effect of ferroptosis on the ultrastructure of granulosa cells， the high pressure freezing-freezing substitution method （HPF-FS） and conventional chemical fixation method （CF） were used to study the ultrastructures of bovine and porcine granulosa cells induced by Erastin. The results showed that， after induction by Erastin for 48 h， the proliferation of both bovine granulosa cells and porcine granulosa cells were significantly inhibited， the ATP contents were significantly decreased， and the ROS levels were significantly increased when compared with the control group， all of these indicated ferroptosis in granulosa cells. Electron microscopy observation indicated that vacuoles appeared in the cells， mitochondria size were significantly decreased， cristae broken or even disappeared of granulosa cells in both treatment groups； the endoplasmic reticulum of the control porcine granulosa cells were more abundant than the treatment group. The maintenance of the ultrastructure of the cells by HPF-FS were better than CF in both cell morphology and membrane boundary. Compared with CF， the structures of mitochondria， golgi apparatus， autophagy and nucleus were clear and intact， which were closer to the transient physiological state of cells. Therefore， the technology of HPF-FS can more realistically and comprehensively display the ultrastructure of cells， which can providing more ideas and references for further research on the apoptosis regulation mechanism of granulosa cells or the related research on follicle development.

The microbial communities has an important impact on the macro nature of the environment. However， the characteristics of high-dimensional， complex and sparse microbial data also pose new challenges for understanding the relationship between microorganisms and ecological environment. The development of machine learning and the popularization of the application of the second generation DNA sequencing technology provided a new solution to this problem. In this study， soil microbiome and dissolved organic carbon （DOC） data of 308 samples from plant litter decomposition experiments for 44 days were used， and 1 709 operational taxonomic units （OTU） of bacteria and microorganisms were used as features to build 12 commonly used machine learning models. Embedding method， packaging method and embedd-packaging fusion method were used for feature selection， and gradient boosting decision tree （GBDT） was selected as the optimal model for parameter optimization. The model adopted root mean square error， mean absolute error and linear goodness of fit was used as evaluation indexes. The results showed that， the feature selection reduced the data dimension and improved the model accuracy. In the simulation experiment， the embedding-packaging fusion method performs was the best in the application model. The prediction model of dissolved organic carbon was constructed by combining the embedding and packaging fusion method with gradient boosting decision tree， and the validity of the model was verified by experiments. The results provided a new way to estimate dissolved organic carbon using machine learning method based on bacterial and microbial data.