Improving (Q)SAR predictions by examining bias in the selection of compounds for experimental testing

ABSTRACT

Existing data on structures and biological activities are limited and distributed unevenly across distinct molecular targets and chemical compounds. The question arises if these data represent an unbiased sample of the general population of chemical-biological interactions. To answer this question, we analyzed ChEMBL data for 87,583 molecules tested against 919 protein targets using supervised and unsupervised approaches. Hierarchical clustering of the Murcko frameworks generated using Chemistry Development Toolkit showed that the available data form a big diffuse cloud without apparent structure. In contrast hereto, PASS-based classifiers allowed prediction whether the compound had been tested against the particular molecular target, despite whether it was active or not. Thus, one may conclude that the selection of chemical compounds for testing against specific targets is biased, probably due to the influence of prior knowledge. We assessed the possibility to improve (Q)SAR predictions using this fact: PASS prediction of the interaction with the particular target for compounds predicted as tested against the target has significantly higher accuracy than for those predicted as untested (average ROC AUC are about 0.87 and 0.75, respectively). Thus, considering the existing bias in the data of the training set may increase the performance of virtual screening.     

Spectrophotometric resolution of ternary mixtures of pseudoephedrine hydrochloride, dextromethorphan hydrobromide, and sodium benzoate in syrups using wavelength selection by net analyte signals calculated with hybrid linear analysis

Hybrid linear analysis (HLA), as a recent factor-based multivariate calibration technique, was applied for the spectrophotometric determination of ternary mixtures of pseudoephedrine hydrochloride (PSU), dextromethorphan hydrobromide (DXT), and sodium benzoate (BNZ). The utilized HLA was assisted by a wavelength selection procedure which was based on the calculation of the net analyte signal (NAS) regression plot in any considered wavelengths window for each test sample, in addition to a moving window strategy for searching the region with maximum linearity of NAS regression plot (minimum error indicator (EI)). HLA was applied because it was simpler to adapt to the NAS regression plot methodology, and also used less factors than partial least squares (PLS). An orthogonal array design was applied for formation of calibration and prediction sets in the concentration ranges 0-7500 μmol L⁻¹ for PSU, 0-300 μmol L⁻¹ for DXT, and 0-1400 μmol L⁻¹ for BNZ. The method had the ability to select wavelength regions that minimize the effect of non-linearity of the spectral data, in addition to that of non-modeled interferences. The application of the selected wavelength regions improved the obtained relative standard error of predictions for PSU, DXT, and BNZ, respectively, from 5.24, 8.67, and 5.48% to 2.19, 5.21, and 3.62% (using lower number of factors). To check the ability of the proposed method in selection of linear regions of spectra, a test for detecting non-linear regions of spectral data in multivariate spectroscopic assays was also described. Additives in the commercial syrup samples did not interfere with their determinations. The method was successfully applied for the determination of pseudoephedrine HCl, dextromethorphan HBr, and sodium benzoate in cough suppressant syrup samples.     

Packings and stationary phases for biopolymer separations by HPLC

Summary Packings and stationary phases applied to high resolution separations of proteins, enzymes, and nucleic acids must satisfy a series of distinct criteria that are different from those usually required by HPLC of low molecular weight non-biologically active analytes. These requirements have been met through substantial improvements in classical gel media together with novel developments in silica supports, and have led to a family of products with tailor-made and reproducible properties. Supports consisting of cross-linked organic gels, and inorganic materials (mostly silicas) are now available with graduated particle sizes, pore sizes, porosities and surface areas as well as non-porous beads. A whole range of stationary phases, such as reversed phase, hydrophobic interaction, ion exchanger and affinity packings, were designed for application as chemical sensors for biopolymer recognition in adsorptive chromatography. The phase systems are operated in the gradient mode, giving high resolution and high peak capacities. In addition, aqueous liquid-liquid partitioning systems have been developed for the fractionation of proteins and nucleic acids. Size exclusion media complete the set of HPLC variants enabling a discrimination of proteins according to their size and shape in an isocratic elution mode. Basically, protein purification and isolation is a multistage process where-by the HPLC variants are combined in a logistic sequence, utilizing the different selectivities of the phase systems and thus maximising resolution, speed and throughput.     

一种新型极性脲丙基-C30固定相的制备和性能评价

以3-脲丙基三甲氧基硅烷为偶联剂，制备了一种新型极性脲丙基-C30（TPU-C30）反相色谱固定相。采用扫描电子显微镜、元素分析、红外光谱和热分析等对该固定相进行表征。结果表明，TPU-C30固定相已成功制备，连续制备3次固定相，其元素含量的相对偏差均小于5%，说明该合成工艺重复性良好。以不同极性、位置异构、碱性化合物为溶质探针，以传统的C18色谱柱与C30色谱柱为参比，对制备的固定相的色谱性能进行了研究。研究结果表明，TPU-C30固定相具有不同于传统C18柱、C30柱的选择性和更优的择形性，明显改善了碱性物质的峰形，其具有广阔的应用空间。     

Single-Round DNA Aptamer Selection by Combined Use of Capillary Electrophoresis and Next Generation Sequencing: An Aptaomics Approach for Identifying Unique Functional Protein-Binding DNA Aptamers

In DNA aptamer selection, existing methods do not discriminate aptamer sequences based on their binding affinity and function and the reproducibility of the selection is often poor, even for the selection of well-known aptamers like those that bind the commonly used model protein thrombin. In the present study, a novel single-round selection method (SR-CE selection) was developed by combining capillary electrophoresis (CE) with next generation sequencing. Using SR-CE selection, a successful semi-quantitative and semi-comprehensive aptamer selection for thrombin was demonstrated with high reproducibility for the first time. Selection rules based on dissociation equilibria and kinetics were devised to obtain families of analogous sequences. Selected sequences of the same family were shown to bind thrombin with high affinity. Furthermore, data acquired from SR-CE selection was mined by creating sub-libraries that were categorized by the functionality of the aptamers (e. g., pre-organized aptamers versus structure-induced aptamers). Using this approach, a novel fluorescent molecular recognition sensor for thrombin with nanomolar detection limits was discovered. Thus, in this proof-of-concept report, we have demonstrated the potential of a “DNA Aptaomics” approach to systematically design functional aptamers as well as to obtain high affinity aptamers.     

Random frog: An efficient reversible jump Markov Chain Monte Carlo-like approach for variable selection with applications to gene selection and disease classification

The identification of disease-relevant genes represents a challenge in microarray-based disease diagnosis where the sample size is often limited. Among established methods, reversible jump Markov Chain Monte Carlo (RJMCMC) methods have proven to be quite promising for variable selection. However, the design and application of an RJMCMC algorithm requires, for example, special criteria for prior distributions. Also, the simulation from joint posterior distributions of models is computationally extensive, and may even be mathematically intractable. These disadvantages may limit the applications of RJMCMC algorithms. Therefore, the development of algorithms that possess the advantages of RJMCMC methods and are also efficient and easy to follow for selecting disease-associated genes is required. Here we report a RJMCMC-like method, called random frog that possesses the advantages of RJMCMC methods and is much easier to implement. Using the colon and the estrogen gene expression datasets, we show that random frog is effective in identifying discriminating genes. The top 2 ranked genes for colon and estrogen are Z50753, U00968, and Y10871_at, Z22536_at, respectively. (The source codes with GNU General Public License Version 2.0 are freely available to non-commercial users at: http://code.google.com/p/randomfrog/.)     

A maximum-efficiency-first multi-path route selection strategy for optical burst switching networks

In this paper, the impact of a path selection on other existing paths in optical burst switching (OBS) networks is studied by analyzing the contention among different traffic streams and the interaction between the route selection and traffic load balance. The results show that there exists a mutual reinforcement interaction among the traffic load of a path, the path burst loss ratio and the contention ability of the path when burst loss ratio based multi-path selection strategies are adopted, which may increase the unbalance of traffic and lead to severe congestion further. A maximum-efficiency-first multi-path selection strategy, which considers the performance of the burst flows and the impact of a path selection on existing OBS paths at the same time by a combined metric of route efficiency, is proposed to maximize the utility of the burst flows and minimize the increment of lost throughput on the path. The performance of the proposed multi-path selection strategy is evaluated through simulation. The results show that the presented strategy obviously outperforms the least burst loss ratio strategy and shortest path first strategy in terms of the burst loss ratio in the practical unbalanced background traffic, especially when the network is heavily loaded.     

Monotonicity of certain integrals involving gamma distributions and their applications in multiple comparisons

Barlow and Gupta (1969) and Alam (1970) studied the monotonicity of two integrals, involving gamma distributions, that arise in certain ranking and selection problems. In this paper, we shall unify their results by studying the monotonicity of two generalized versions of integrals considered by them. We will also provide applications of derived results in study of certain multiple comparison procedures.