不相交主成分分析(PCA)和遗传算法(GA)用于差异表达基因的识别

建立了一种基于不相交主成分分析(Disjoint PCA)和遗传算法(GA)的特征变量选择方法, 并用于从基因表达谱(Gene expression profiles)数据中识别差异表达的基因. 在该方法中, 用不相交主成分分析评估基因组在区分两类不同样品时的区分能力; 用GA寻找区分能力最强的基因组; 所识别基因的偶然相关性用统计方法评估. 由于该方法考虑了基因间的协同作用更接近于基因的生物过程, 从而使所识别的基因具有更好的差异表达能力. 将该方法应用于肝细胞癌(HCC)样品的基因芯片数据分析, 结果表明, 所识别的基因具有较强的区分能力, 优于常用的基因芯片显著性分析(Significance analysis of microarrays, SAM)方法.     

对直接标准化算法的改进及其应用

由于各种仪器之间存在差异,主机上建立的定量模型用于从机会导致预测结果出现较大偏差。目前主要通过有标样方法和无标样方法来减小预测偏差。该文对现有标样方法中的直接标准化算法进行改进,在转移矩阵的建立过程中,对从仪器数据矩阵进行主成分分解,以预测均方差为判定标准,确定最终的转移矩阵。并以玉米和烟草数据为对象,测试了该法的有效性。玉米样品含有2种成分:水分和蛋白质;烟草样品含有4种成分:还原糖、总糖、总氮和总碱。结果表明,对于玉米样品中的2种成分,采用改进的方法可显著提高预测的准确度;对于烟草中的4种成分而言,采用改进的方法可获得稳健的预测结果。     

Recognition of big data mixed Raman spectra based on deep learning with smartphone as Raman analyzer

Raman spectral detection has emerged as a powerful analytical technique due to the advantages of fast acquisition, non-invasion, and low cost. The on-site application is highly dependent on Raman automatic analysis algorithm. However, current Raman algorithm research mainly focuses on small sample Raman spectroscopy (RS) identification with defects of low accuracy and detection rate. It is also difficult to realize rapid RS measurement under big data. In this paper, rapid recognition of mixtures in complex environments was realized by establishing a fast Raman analysis model based on deep learning through data training, self-learning, and parameter optimization. The cloud network architecture was proposed to apply deep learning to real-time detection using Smartphone-based Raman devices. This research solves the technical problems about mixture recognition under big data and thus could be used as a new method for fast and field RS detection in complex environments.     

一种基于最小二乘法的石英晶体微分天平的谐振频率检测算法

阻抗型石英晶体微天平,具有适用范围广、获取参数多等优点,但通过阻抗分析方法获取石英晶体谐振频率耗时较长,实时性差。 针对这一问题,对采用最小二乘法进行二次拟合的计算过程进行了深入的分析,对其中的关键步骤进行了改进,提出了一种适用于石英晶体谐振频率测量的快速检测算法,通过对测得数据的坐标系进行调整,降低了拟合过程中的数据运算量。 经实验验证,与传统拟合算法相比,该算法计算谐振频率的效率提升4倍,并且得到了更高的计算精度,具有很高的实用价值。     

New method for parallel computation of Hessian matrix of conformational energy function in internal coordinates

A new algorithm for parallel calculation of the second derivatives (Hessian) of the conformational energy function of biomolecules in internal coordinates is proposed. The basic scheme of this algorithm is the division of the entire calculation of the Hessian matrix (called "task") into subtasks and the optimization of the assignment of processors to each subtask by considering both the load balancing and reduction of the communication cost. A genetic algorithm is used for this optimization considering the dependencies between subtasks. We applied this method to a glutaminyl transfer RNA (Gln-tRNA) molecule for which the scalability of our previously developed parallel algorithm was significantly decreased when the large number of processors was used. The speedup for the calculation was 32.6 times with 60 processors, which is considerably better than the speedup for our previously reported parallel algorithm. The elapsed time for the calculation of subtasks, data sending, and data receiving was analyzed, and the effect of the optimization using the genetic algorithm is discussed.     

Tucker1 model algorithms for fast solutions to large PARAFAC problems

We describe a method of performing trilinear analysis on large data sets using a modification of the PARAFAC‐ALS algorithm. Our method iteratively decomposes the data matrix into a core matrix and three loading matrices based on the Tucker1 model. The algorithm is particularly useful for data sets that are too large to upload into a computer's main memory. While the performance advantage in utilizing our algorithm is dependent on the number of data elements and dimensions of the data array, we have seen a significant performance improvement over operating PARAFAC‐ALS on the full data set. In one case of data comprising hyperspectral images from a confocal microscope, our method of analysis was approximately 60 times faster than operating on the full data set, while obtaining essentially equivalent results. Copyright © 2008 by John Wiley & Sons, Ltd.     

On One Method of Parameter Estimation in Chemical Kinetics Using Spectra with Unknown Spectral Components

The method of successive estimation of regression parameters, which is widely used in nonlinear regression analysis, is applied to obtain kinetic information from spectral data for the case when the spectra of individual components are unknown. Using a model example with a two-step successive reaction, the reliability of the algorithm is demonstrated. To compare the proposed method with other known methods for estimating kinetic parameters literature data are used. All simulations were done using a new software for nonlinear regression analysis: FITTER. The proposed approach is especially useful when the spectra of reaction components are unknown and when formal calibration methods do not provide desirable accuracy.     

通用模拟退火用于稳健多元分析校正

模拟退火是一种全局优化算法，具有跨越局部最优点的机制，最小一乘是一种较常用的最小二乘更为稳健的优化准则，更适用于可能偏离正态分布的实际数据集，本文探讨了用最小一乘为准则并利用模拟退火方法同时测定多组分体系的可能性。应用于2－3组分药物体系分析，获得了满意的结果，本文还探讨了改变步长提高模拟退火算法优化精度的方法。     

A novel method for human gender classification using Raman spectroscopy of fingernail clippings

This paper illustrates a novel method for human gender classification by measuring the Raman spectrum of fingernail clippings. As Raman spectroscopy reveals the characteristics of vibrational frequencies of the fingernails, it provides unique chemical fingerprints that can be used to describe the molecular structure differences of fingernail between males and females. As the differences of Raman spectra of human fingernails are very subtle, they are enhanced by using a pattern recognition method. In the present study, a combination algorithm of principal component analysis (PCA) and support vector machines (SVM) was implemented to perform the data classification. This combined algorithm provides a classification accuracy of up to 90%. The success of this present method may be used as an alternative rapid tool to identify human gender in forensic applications.     

Evolving projection analysis of multicomponent mixtures

The application of evolving projection analysis (EPA) to second-order bilinear data sets consisting of more than two components is described. EPA is a method for rank analysis of ordered data matrices where the components appear sequentially as a function of time or some other ordinal variable. It was found that extension of the method to mixtures of more than two components was best accomplished using principal components analysis to preprocess the data. The algorithm is demonstrated using simulated four-component chromatographic data, and experimental data from liquid chromatography (three- and four-component mixtures) and a spectrophotometric titration (four components), both employing UV-visible diode array detection.