基于改进排序遗传算法的径向基函数神经网络色谱峰解析

构造了以塔板模型为基函数的径向函数神经网络（P－RBFNN），为了使P－RBFNN具有结构重组能力，又在网络学习算法中引入鲁棒（Rubust)和随机全局最优的两阶段排序遗传算法：结构学习和进化。P－RBFNN结合改进的排序遗传算法很适合组分数未知的色谱（含重叠）峰解析。     

基于稀疏模型的气相色谱-质谱数据解析及其在严重重叠峰解析中的应用

提出一种气相色谱-质谱（GC-MS）数据解析算法。以色谱峰顶点处的质谱作为待测谱,在谱库中检索一定量相关参考谱,求解关于各纯组分色谱响应值的方程。质谱检索采取分步策略,先利用质谱碎片规律建立高速索引进行粗选,然后使用强峰高概率出峰准则和耐挤压性准则排除无关质谱。为求解色谱响应值方程,提出基于稀疏模型的回归算法,相比传统算法,稀疏算法利于提取待测谱的主要结构,避免"过拟合"。实验结果表明,该质谱检索算法具有较高的精度和规模较小的剩余参考谱集,而所提稀疏模型算法可有效解析严重重叠峰。该算法可作为GC-MS重叠峰解析,特别是严重重叠峰解析的一种有效解决方案。     

基于色谱峰形优劣的代谢组学峰检测参数优化算法比较

非靶向代谢组学数据预处理的关键步骤之一为峰检测（Peak picking）,在高分辨质谱的峰检测过程中应用最广泛的算法为基于连续小波变换的cent Wave算法。本研究通过代谢物标准品和尿液两个数据集,结合优良峰形色谱峰比例、可信色谱峰比例和可重复色谱峰比例3个评价指标对IPO(Isotopologue parametersoptimization)和cent Wave Sweep两种cent Wave参数优化算法进行了全面比较。为了快速准确地对色谱峰形优劣进行区分,比较了随机森林（Random forest）、自适应提升（Adaboost）和梯度提升树（Gradient boosting decisiontree） 3种集成学习算法在区分色谱峰形优劣方面的性能。根据准确度和衡量二分类模型精确度的F1分数,选择随机森林建立区分模型（准确度93.5%, F1分数0.938）。研究结果表明,相比于XCMS Online推荐的参数,采用IPO和cent Wave Sweep进行参数优化后,不同数据集的可信色谱峰比例和可重复色谱峰比例均得到了提高,取得了较好的优化效果;但是,对于不同数据集的优...     

小波变换用于重叠色谱峰组分信息的提取

提出了一个小波变换用于提取重叠色谱中组分信息的理论公式，并将它应用于三组分和五组分重叠色谱实验数据的解析。结果表明：在定定分离度范围内，当相邻色谱峰峰宽相近时，根据此公式的计算结果，可以将各组分的信息从重叠色谱峰信号中提取出来。洒工作对小波变换用于重叠峰的解析具有重要意义。     

非对称色谱峰的拟合与重叠峰的拟合解析

 ]利用色谱保留时间tR与色谱峰形模型参数tG、σ、τ间的线性关系、预测了气相色谱填充柱上非对称的和重迭的色谱峰的有关参数，并据此对这两类色谱峰进行了拟合和解析。与常用重叠峰解析方法如中垂线法，三角形法和切线法解析结果的对比证明，曲线拟合法在这些方法中具有最好的准确度。用曲线拟合法进行实际定量的结果表明，对于非对称峰，其定量的相对误差小于0.5％，而对于重叠峰的定量误差则在10％以内。     

完全重叠色谱峰的二维微分色谱解析算法

利用完全重叠的色谱峰中同时消失的两组分的浓度变化速率差异，通过二维数据微分色谱法获取被掩盖组份的近似光谱，以解析完全重叠的色谱体系，将该法应用于维生素Ｃ和菸酸两组份完全重叠的二维色谱之分辨，结果令人满意，分辨光谱与标准光谱重合甚好。     

模差渐进投影法解析重叠色谱峰

提出了一种鉴别重叠色谱峰类型的化学计量学方法——模差渐进投影(MDEP)法。该方法利用二维色谱数据中相邻两个色谱流出点的光谱矢量模的和与该两个矢量和的模之差,建立模差-时间(dm-t)曲线判断非纯色谱峰的纯组分区,同时判定第二组分色谱流出点和倒数第二组分色谱结束点。分别以纯组分区的光谱矢量建立正交阵对该色谱峰进行投影运算,得到投影后转换色谱。该转换色谱消除了纯组分区所对应化合物的色谱和光谱信息。对转换色谱重复上述操作至整个色谱峰的模差-时间曲线均为0时终止,得到重叠色谱峰各组分的色谱流出点和结束点。利用上述结果,采用不包含待测组分光谱信息的该重叠峰二维色谱数据建立正交矩阵对该色谱峰进行投影运算,获得待测组分的相似色谱图。通过比对分析即可得到该组分在重叠色谱峰中的光谱和色谱信息。依次操作,即可获得重叠色谱峰中各组分的光谱和色谱信息。采用模差渐近投影法对4组分模拟重叠二维色谱峰及杜香挥发油GC-MS的重叠色谱峰进行了解析,方法简便、结果准确。     

一阶和二阶导数相综合的色谱峰检测法

本文提出了一种介于一阶和二阶导数之间的色谱峰检测新方法，它能排除基线漂移的影响，具有抗噪声干扰能力，能判别重叠峰，还能够得色谱峰上各拐点，的位置并发现肩峰型重叠峰。该方法只需要设置一个起动参数即可自动进行峰检测，具有很强的自适应性和健壮性，极大地方便了用户。     

严重重叠的色谱峰对的解析

本文根据指数修正的高斯函数模型(EMG),并利用回归分析提出了一种解析严重重叠的色谱峰对的方法。此法可适用于相对峰谷在50％～95％的严重重叠的色谱峰对的解析。     

基于积分渐进展开的气相色谱重叠峰快速解析法

论文提出用积分渐进展开解析气相色谱重叠峰,该方法有3个主要步骤：首先将谷峰或肩峰分成两个积分区域,得到一个子区域的积分方程和一个重叠峰面积的代数方程;然后用数值积分求出这两个方程计算中所需要的峰面积,再用积分渐进公式将积分方程展开成代数方程;最后,将这两个方程与峰高约束方程联立后,得到一个非线性代数方程组,用Gauss-Seidel迭代可以快速求解方程组,方程收敛的最大迭代次数不超过20次。仿真和实验结果表明,解析的峰高和峰面积误差均很小,峰面积最大误差低于6.44%,峰高的最大误差约为6.80%。由于该算法精度高,效率高,所以这个方法可以用于气相色谱重叠峰和一般色谱峰的实时在线解析。     

Nested genetic algorithm for resolving overlapping spectra

A nested genetic algorithm, including a genetic parameter level and a genetic implemented level, has been proposed and applied for resolving simulated overlapping spectra. Parameters of genetic algorithms (GA) were optimized by use of the genetic parameter level. The number of overlapping peaks was, moreover, detected simultaneously. Parameters of individual peaks in multiplets were computed by use of the genetic implemented level. It is obvious that the parameters of GA can be optimized and the number of overlapping peaks can be detected by itself. The optimization results are less affected than in traditional curve-fitting by the initial values of the parameters of the overlapping bands. Consequently, the nested genetic algorithm is superior to the curve-fitting technique for resolution of overlapping peaks.     

Combined Kalman filtering and steepest descent minimization for quantitative analysis of unresolved misaligned chromatographic peaks. The resolution of alternariol and altenuisol mycotoxins in mixtures by HPLC

Summary The potential of a computational approach for the quantitative resolution of seriously overlapping chromatographic peaks when there is loss of collinearity between the pure component peaks and the mixture peak has been explored. The program makes iterative use of the Kalman filter algorithm for resolving the mixture peak with the component peaks aligned according to some values of the position parameters, and of a steepest descent minimization procedure to find the optimal alignment. This combined procedure has been applied to the quantitive resolution of the HPLC chromatograms of alternariol and altenuisol mycotoxins in synthetic mixtures and in real samples.     

Resolution of Overlapping Chromatographic Peaks Using a Genetic Algorithm

ABSTRACT

A genetic algorithm for resolution of overlapping chromatographic peaks (GAROCP) using real-number coding, non-uniform mutation and arithmetical crossover methods is described in this paper. It was applied to resolution of highly overlapped multicomponent high-performance liquid chromatographic peaks by fitting experimental chromatogram to the exponentially modified Gaussian (EMG) model. The genetic algorithm was used to find the minimum of fitting error to optimize the parameters in the EMG functions which determine the shape and area of each peak. The applicability of the method was investigated with both simulated signals calculated by EMG functions and experimental multicomponent overlapping chromatograms.     

改进遗传算法用于未知组分数的重叠色谱峰的解析

按照指数修正的Gauss卷积色谱峰模型，构造了结合模拟退火及变长染色体的改进遗传算法，并结合爬山法，可以方便快捷地进行数目未知的重叠色谱峰的精确解析；此法成功地应用于大鼠脑微透析液氨基酸定量分析中，解析结果良好。     

Prediction of inter-residue contacts map based on genetic algorithm optimized radial basis function neural network and binary input encoding scheme

Summary Inter-residue contacts map prediction is one of the most important intermediate steps to the protein folding problem. In this paper, we focus on the problem of protein inter-residue contacts map prediction based on neural network technique. Firstly, we use a genetic algorithm (GA) to optimize the radial basis function widths and hidden centers of a radial basis function neural network (RBFNN), then a novel binary encoding scheme is employed to train the network for the purpose of learning and predicting the inter-residue contacts patterns of protein sequences got from the protein data bank (PDB). The experimental evidence indicates the utility of our proposed encoding strategy and GA optimized RBFNN. Moreover, the simulation results demonstrate that the network got a better performance for these proteins, whose residue length falls into the area of (100, 300), and the predicted accuracy with a contact threshold of 7 Å scores higher than the other 3 values with 5, 6, and 8 Å .     

Translation Modification Iteration for Resolution and Quantification of Overlapping Chromatographic Peaks

A novel algorithm, translation modification iteration (TMI), is proposed for resolving overlapping chromatographic peaks. By this method, a series of similar peaks are obtained from the initial construction of single component peaks and approach the profiles of real single component peaks by iteration. Both simulated and experimental data are investigated with TMI, consequently overlapping peaks can be resolved into reasonable single component peaks easily and efficiently. Quantitative analysis is also successfully achieved in both simulated and experimental data, and the quantitative results obtained from TMI are superior to those obtained from perpendicular drop (PD) and tangent skim (TS) methods. Without too much time devoted to the procedure, TMI is a simple and practicable method for resolution and quantification of overlapping chromatographic peaks.

     

基于遗传算法的色谱指纹峰配对识别方法

指纹峰配对识别是色谱指纹图谱分析中的关键环节之一，本文提出一种基于遗传算法的色谱指纹峰配对识别方法。该法根据对照色谱指纹图谱的峰分布特性初选出若干标定蜂，将其存入一个候选标定蜂库；同时根据这些候选标定峰从待测指纹图谱中选出相应的候选标定峰，也存入候选标定峰库；再用遗传算法从库中选取一组标定峰用于校正待测指纹图谱中各峰的峰位，并自动识别出与对照色谱指纹图谱相对应的各指纹峰。仿真实验及实际分析实验结果均表明，该法识别指纹峰准确可靠，可用于色谱指纹图谱相似度的快速自动计算。     

免疫算法: 一种新的多组分重叠色谱信号的解析方法

色谱及其联用技术的快速发展使其成为复杂样品分析的重要手段,但在复杂样品分析中仍难以避免多组分色谱信号的重叠。化学计量学利用数学方法对复杂重叠信号进行解析,为复杂样品的色谱分析提供了新的手段。本文对免疫算法研究的近期进展进行了总结,并以具体的实例介绍了其在多组分重叠色谱信号解析方面的应用。