A Method for Clustering and Screening of Long-dimensional Chemical Data Based on Fingerprints and Similarity Measurements

A method for the treatment of long-dimensional chemical data arrays is presented in this work with the aim of maximising classification models. The method is based on the construction of fingerprints and the subsequent generation of a similarity matrix. The similarity calculation has been modified through a scaling process to take into account different significance shown by the variables. The method was applied to spectral measurements of wines and several aspects were studied, namely: threshold considered in the construction of fingerprints and patterns, weighting factor used for scaling, normalisation method, etc. The application of both Principal Components Analysis and Soft-Independent Modelling of Class Analogies to the similarity matrices gave better classifications of the information than those obtained using original data.     

Introduction of the conditional correlated Bernoulli model of similarity value distributions and its application to the prospective prediction of fingerprint search performance

A statistical approach named the conditional correlated Bernoulli model is introduced for modeling of similarity scores and predicting the potential of fingerprint search calculations to identify active compounds. Fingerprint features are rationalized as dependent Bernoulli variables and conditional distributions of Tanimoto similarity values of database compounds given a reference molecule are assessed. The conditional correlated Bernoulli model is utilized in the context of virtual screening to estimate the position of a compound obtaining a certain similarity value in a database ranking. Through the generation of receiver operating characteristic curves from cumulative distribution functions of conditional similarity values for known active and random database compounds, one can predict how successful a fingerprint search might be. The comparison of curves for different fingerprints makes it possible to identify fingerprints that are most likely to identify new active molecules in a database search given a set of known reference molecules.     

不同产地白芷药材高效液相色谱指纹图谱研究

本文采用高效液相色谱-二极管阵列检测器(HPLC-DAD)法建立中药白芷的指纹图谱.应用化学计量学中两种不同的模式识别方法(主成分分析法和系统聚类分析法)对实验数据进行处理,以找出来自三个不同产地30个中药白芷样品间的相似性及差异性.两种模式识别方法均能成功地按样品的来源将不同产地的样品正确分类.建立了不同产地中药白芷的识别方法,该方法能有效地控制中药白芷的质量,并能为其它中药产品的化学模式识别提供参考.     

Multifingerprint based similarity searches for targeted class compound selection

Molecular fingerprints are widely used for similarity-based virtual screening in drug discovery projects. In this paper we discuss the performance and the complementarity of nine two-dimensional fingerprints (Daylight, Unity, AlFi, Hologram, CATS, TRUST, Molprint 2D, ChemGPS, and ALOGP) in retrieving active molecules by similarity searching against a set of query compounds. For this purpose, we used biological data from HTS screening campaigns of four protein families (GPCRs, kinases, ion channels, and proteases). We have established threshold values for the similarity index (Tanimoto index) to be used as starting points for similarity searches. Based on the complementarities between the selections made by using different fingerprints we propose a multifingerprint approach as an efficient tool to balance the strengths and weaknesses of various fingerprints.     

Mathematical correction for fingerprint similarity measures to improve chemical retrieval

In many modern chemoinformatics systems, molecules are represented by long binary fingerprint vectors recording the presence or absence of particular features or substructures, such as labeled paths or trees, in the molecular graphs. These long fingerprints are often compressed to much shorter fingerprints using a simple modulo operation. As the length of the fingerprints decreases, their typical density and overlap tend to increase, and so does any similarity measure based on overlap, such as the widely used Tanimoto similarity. Here we show that this correlation between shorter fingerprints and higher similarity can be thought of as a systematic error introduced by the fingerprint folding algorithm and that this systematic error can be corrected mathematically. More precisely, given two molecules and their compressed fingerprints of a given length, we show how a better estimate of their uncompressed overlap, hence of their similarity, can be derived to correct for this bias. We show how the correction can be implemented not only for the Tanimoto measure but also for all other commonly used measures. Experiments on various data sets and fingerprint sizes demonstrate how, with a negligible computational overhead, the correction noticeably improves the sensitivity and specificity of chemical retrieval.     

New approach on similarity analysis of chromatographic fingerprint of herbal medicine

A new approach to the construction and similarity analysis of chromatographic fingerprint for herbal medicine is presented in this paper. Samples of chuanxiong, a herbal medicine for headache, from three producing areas of China were used to evaluate the utility of this study. The samples were analyzed with high-performance liquid chromatography (HPLC) and the peak areas of the chromatograms were used to construct the fingerprints of the herbal medicines. A vector of differences was defined between the two fingerprints. The scalar mean of the difference vector was taken as a statistic and both the t-test and Bayesian hypothesis testing were implemented to provide a one-to-one comparison of the fingerprints. Compared with principal component analysis (PCA), correlation coefficient and vector cosine, the new method offers a better differentiation of the similarity or difference between the fingerprints from same sample of chuanxiong. When the new method was used in the similarity analysis of the fingerprints of chuanxiong from different production areas, a clear-cut signature was obtained that reveals the significant difference between them.     

NMR spectroscopy evaluation of direct relationship between soils and molecular composition of red wines from Aglianico grapes

¹H NMR spectroscopy was employed to investigate the molecular quality of Aglianico red wines from the Campania region of Italy. The wines were obtained from three different Aglianico vineyards characterized by different microclimatic and pedological properties. In order to reach an objective evaluation of “terroir” influence on wine quality, grapes were subjected to the same winemaking procedures. The careful subtraction of water and ethanol signals from NMR spectra allowed to statistically recognize the metabolites to be employed in multivariate statistical methods: Principal Component Analysis (PCA), Discriminant Analysis (DA) and Hierarchical Clustering Analysis (HCA). The three wines were differentiated from each other by six metabolites: α-hydroxyisobutyrate, lactic acid, succinic acid, glycerol, α-fructose and β-d-glucuronic acid. All multivariate analyses confirmed that the differentiation among the wines were related to micro-climate, and carbonate, clay, and organic matter content of soils. Additionally, the wine discrimination ability of NMR spectroscopy combined with chemometric methods, was proved when commercial Aglianico wines, deriving from different soils, were shown to be statistically different from the studied wines. Our findings indicate that multivariate statistical elaboration of NMR spectra of wines is a fast and accurate method to evaluate the molecular quality of wines, underlining the objective relation with terroir.     

Brand-dependent volatile fingerprinting of Italian wines from Valpolicella

Head-space solid-phase microextraction (HS-SPME) coupled to gas-chromatography-mass spectrometry was developed and applied to obtain the volatile aromatic fingerprints of three typical Italian wines, Valpolicella, Amarone and Recioto, all produced in the restricted geographical area of Valpolicella (Veneto, Italy) with the same grape cultivars within the regulations of a rigid disciplinary of production. Differences between the three typologies are mainly linked to the different withering times to which grapes are subjected before vinification, which strongly influences the concentration and the development of volatile aroma compounds. A total of 22 different wines (7 Valpolicella, 10 Amarone and 5 Recioto) were characterised in terms of aromatic volatile profile with the aim to distinguish the different products and to evaluate the possibility to differentiate the same product from different brands. For the chemometric evaluation of the data one-way analysis of variance (ANOVA), principal component analysis (PCA) and hierarchical cluster analysis (HCA) were tested. All the chemometric tools employed allow to differentiate between the three products. More intriguing is the ability of the chemometric approach to differentiate between the same product (Amarone, Recioto) from different winery, thus showing the potential of this approach to characterize the brand-dependent typicality of wines, which is usually related to subtle technological differences which nevertheless have strong influences on the organoleptic characteristics of the products.     

化学环境编码与分子相似度计算

分子相似度方法主要用于药物分子设计中先导化合物的选取及化合物物理、化学性质的预测,这种方法所依据的原理是：相似的化合物结构将具有相似的化学及物理性质［‘,’1．将已知具有某种性质的化合物的结构（常称为探针化合物）与诸多化合物进行比较,由此得到与之相似的化合物,这种邻近化合物有可能在分子设计中做先导化合物．近年来已报道过多种化合物相似度的计算方法,可粗略的归为以下几类：（l）结构碎片法［‘’‘j;（2）拓扑指数法卜·“;（3）量子化学方法［’,’j．本文提出一种新的化学环境编码方法,该方法有别于结构碎片…     

Combination of several mass spectrometry ionization modes: a multiblock analysis for a rapid characterization of the red wine polyphenolic composition

In the present study, direct flow injection mass spectrometry was investigated for rapid characterization of the polyphenolic composition of red wines. Atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) (in both positive and negative ion modes) have been simultaneously used for a more comprehensive analysis of the samples studied. In this way, four mass spectra have been recorded for each wine. Each spectrum was considered as a fingerprint related to the chemical composition. This methodology was applied to a large number of Beaujolais wines from different grades and different vintages.This data set was processed using a chemometrical multiblock analysis, which allowed to synthesize the whole information collected. The results obtained showed that the wine fingerprints address the composition of the main polyphenolic compounds present in the red wines and can discriminate groups of wines showing different polyphenolic compositions. Multiblock analysis appears as a very promising tool to deal with several data tables of multivariate signals in order to define, by combining the whole information, the best operating protocol according to the desired analytical objectives.     

Quality evaluation of fingerprints of herbal medicine with chromatographic data

In this study, local least squares (LLS) and principal component analysis (PCA) were applied to deal with the disturbances in a data set of chromatographic fingerprints after necessary data transformations. It has been demonstrated that PCA with standard normal variate (SNV) transformation of data led to meaningful classification of 33 different Erigeron breviscapus herbal samples. The result was also corroborated by variance squares discriminant method. The quality of herbal objects was further evaluated, and the causes of this fact have been explained from a chemical point of view. At the same time, it implied an idea for qualitative evaluation of the herbal objects with a common class pattern of chromatographic fingerprints.     

The use of high performance liquid chromatography-quadrupole time-of-flight mass spectrometry coupled to advanced data mining and chemometric tools for discrimination and classification of red wines according to their variety

In this study, the potential of high performance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (HPLC–QTOFMS) for metabolomic profiling of red wine samples was examined. Fifty one wines representing three varieties (Cabernet Sauvignon, Merlot, and Pinot Noir) of various geographical origins were sourced from the European and US retail market. To find compounds detected in analyzed samples, an automated compound (feature) extraction algorithm was employed for processing background subtracted single MS data. Stepwise reduction of the data dimensionality was followed by principal component analysis (PCA) and partial least square-discriminant analysis (PLS-DA) which were employed to explore the structure of the data and construct classification models. The validated PLS-DA model based on data recorded in positive ionization mode enabled correct classification of 96% of samples. Determination of molecular formula and tentative identification of marker compound was carried out using accurate mass measurement of full single MS spectra. Additional information was obtained by correlating the fragments obtained by MS/MS accurate mass spectra using the QTOF with collision induced dissociation (CID) of precursor ions.     

Similarity measures based on Gaussian-type promolecular electron density models: alignment of small rigid molecules

Various molecular similarity measures (overlap, Coulomb, kinetic, electrostatic energy) and similarity indices (Carbó, Hodgkin-Richards, Kulczynski, Shape Tanimoto) are applied to the superposition of 3D promolecular electron density (PED) distributions. The original aspect of the paper lies in the consideration of smoothed PEDs, which allow to decrease the number of local solutions to a superposition problem, together with the use of the less common kinetic and electrostatic energy similarity measures. Results are obtained for a family of five rigid endothiapepsin ligands that were already considered in previous applications, based on graph representations of their PED. In the present work, it is observed that the use of smoothed PED and the kinetic similarity measure, together with the Kulczynski or Shape Tanimoto index, performed the best to align molecules of different sizes.     

六味地黄丸的精细指纹图谱分析及模式识别分类研究

采用高效液相色谱建立了六味地黄丸的指纹图谱,对两个厂家的16批产品进行了测定,并结合中药相似度软件和主成分分析法对全指纹图谱和其精细指纹图谱进行了模式识别研究。结果表明,中药相似度软件能够对不同厂家的产品进行区分但也可能造成误判;在主成分分析法的投影图中,两个厂家的产品明显聚为两类,而且不同批次产品的差异也能够显示出来。