Classification of the micro and nanoparticles and biological agents by neural network analysis of the parameters of optical resonance of whispering gallery mode in dielectric microspheres
Classification of the micro and nanoparticles and biological agents by neural network analysis of the parameters of optical resonance of whispering gallery mode in dielectric microspheres
复制标题
通过介电微球回音壁模式光学共振参数的神经网络分析对微米粒子、纳米粒子和生物制剂进行分类
DOI：
10.1117/12.889574
复制
发表时间：
2011
影响因子：
0.7
通讯作者：
A. Ostendorf
中科院分区：
文献类型：
--
作者：
V. Saetchnikov;E. Tcherniavskaia;G. Schweiger;A. Ostendorf
一种利用回音廊模式的光学微腔共振对包括生物分子在内的微米和纳米颗粒进行无标记分析的新技术正在开发中。使用标准微球和特殊生产的微球的各种方法方案已被研究，以进一步发展微生物应用。结果表明，在小于1微瓦的激光功率下，可以检测到最佳几何形状下的光学共振。通过监测耳语走廊模式的光谱漂移，测试了所开发方案的灵敏度。已经使用了乙醇、抗坏血酸、包括白蛋白和盐酸的血液模体、葡萄糖、生物素、C反应蛋白等生物标志物的水溶液，以及细菌和病毒模体(二氧化硅微凝胶和纳米颗粒)。溶液的共振谱结构是一个特定的研究主题。发展了一种用于生物制剂和微纳米颗粒分类的概率神经网络分类器。以光谱位移、展宽、扩散度等光谱参数作为输入参数，建立了微米、纳米颗粒和生物制剂溶液的网络分类器。对于正在研究的探头，分类概率已达到约98%。开发的方法已被证明是一种很有前途的敏感的芯片实验室类型传感器的技术平台，可用于开发不同生物分子的诊断工具，如蛋白质、寡核苷酸、寡糖、脂类、小分子、病毒颗粒、细胞以及不同的实验环境，如蛋白质组学、基因组学、药物发现和膜研究。
A novel technique for the label-free analysis of micro and nanoparticles including biomolecules using optical micro cavity resonance of whispering-gallery-type modes is being developed. Various schemes of the method using both standard and specially produced microspheres have been investigated to make further development for microbial application. It was demonstrated that optical resonance under optimal geometry could be detected under the laser power of less 1 microwatt. The sensitivity of developed schemes has been tested by monitoring the spectral shift of the whispering gallery modes. Water solutions of ethanol, ascorbic acid, blood phantoms including albumin and HCl, glucose, biotin, biomarker like C reactive protein so as bacteria and virus phantoms (gels of silica micro and nanoparticles) have been used. Structure of resonance spectra of the solutions was a specific subject of investigation. Probabilistic neural network classifier for biological agents and micro/nano particles classification has been developed. Several parameters of resonance spectra as spectral shift, broadening, diffuseness and others have been used as input parameters to develop a network classifier for micro and nanoparticles and biological agents in solution. Classification probability of approximately 98% for probes under investigation have been achieved. Developed approach have been demonstrated to be a promising technology platform for sensitive, lab-on-chip type sensor which can be used for development of diagnostic tools for different biological molecules, e.g. proteins, oligonucleotides, oligosaccharides, lipids, small molecules, viral particles, cells as well as in different experimental contexts e.g. proteomics, genomics, drug discovery, and membrane studies.
影响因子：
3.8
作者：
White, Ian M.;Fan, Xudong
通讯作者：
Fan, Xudong