CAREER: Modeling, Estimation and Coding for Biosensor Arrays

研究生物传感器阵列的随机建模和信号处理方面。生物传感器阵列具有时间和成本效益，并且能够在医学，药物发现，防御系统和环境监测中实现令人兴奋的新应用。例如，蛋白质阵列通过同时检测几十种不同的抗原来检测多种病原体感染。另一方面，DNA微阵列能够同时筛选数万甚至数十万个不同的基因序列，揭示关于细胞功能、药物对生物体的影响等的关键信息。然而，为了充分实现生物传感器阵列技术的潜力，必须解决几个关键的研究挑战。感兴趣的生物分子与传感元件之间的分子结合，其使得能够在生物传感器阵列中进行检测，是一个随机过程。实时生物传感器可以对这一过程进行多次时间测量，从而可以观察结合动力学及其稳态的精确表征。研究者的具体目标是：（1）建立随机模型并解决实时生物传感器阵列中的估计问题。本研究涉及解决离散观测系统的随机微分方程建模的参数估计问题。(2)确定生物传感器阵列中估计算法的性能极限，通过最小均方估计误差的下限来表征它们。(3)开发编码策略，提高生物传感器阵列的性能，并研究信号恢复技术，使经济使用的传感资源inherent.The概述的工作的结果，预计将有重大影响的生物传感器阵列的开发和应用，并有望扩大在得克萨斯大学奥斯汀分校的工程专业学生的教育经验。

Project AbstractThe investigator studies stochastic modeling and signal processing aspects of biosensor arrays. The biosensor arrays are time and cost efficient, and enable exciting new applications in medicine, drug discovery, defense systems, and environmental monitoring. Protein arrays, for instance, test for multiple pathogen infections by examining dozens of different antigens at once. DNA microarrays, on the other hand, are capable of screening tens or even hundreds of thousands of different gene sequences at the same time, revealing critical information about the functionality of cells, effects of drugs on organisms, etc. To fully realize the potentials of the biosensor array technology, however, several key research challenges must be addressed.Molecular binding between the bio-molecules of interest and sensing elements, which enables detection in biosensor arrays, is a random process. Real-time biosensors can take multiple temporal measurements of this process, thus allowing observation of the binding kinetics as well as a precise characterization of its steady-state. The investigator specifically aims to: (1) Develop stochastic models and solve the estimation problems in real-time biosensor arrays. This research involves solving parameter estimation problems in discretely observed systems modeled by stochastic differential equations. (2) Determine limits of performance of estimation algorithms in biosensor arrays, characterizing them via lower bounds on the minimum mean-square estimation error. (3) Develop coding strategies that improve the performance of biosensor arrays, and study signal recovery techniques which enable economic use of the sensing resources therein.The results of the outlined work are expected to have a major impact on the development and applications of biosensor arrays, and are expected to broaden the educational experience of engineering students at the University of Texas at Austin.