Similarity Search Approach Research Articles

Ligandbased Virtual screening using Fuzzy Correlation Coefficient

ABSTRACT Selection and identification of a subset of compounds from libraries or databases, which are likely to possess a desired biological activity is the main target of ligand-based virtual screening approaches. The main challenge of such approaches is achieving of high recall of active molecules. In this paper we presented fuzzy correlation coefficients (FCC), which is used as a similarity coefficient. The new approach is based on mutually dependent between molecular features, while most common approaches (Tanimoto, Bayesian and other coefficients) based on mutually independent between features. Our experiments have shown that the new coefficient increases the recall of active molecules in high diversity database compared with other correlation coefficients and Tanimoto. Keywords : Correlation coefficients, fingerprint features, similarity search, similarity coefficients, virtual screening. 1. INTRODUCTION Virtual screening (VS) refers to the use of a computer-based method to process compounds from a library or database of compounds in order to identify and select ones that are likely to possess a desired biological activity, such as the ability to inhibit the action of a particular therapeutic target. Selection of molecules with a virtual screening algorithm should yield a higher proportion of active compounds, as assessed by experiment, relative to a random selection of the same number of molecules [1]. Currently, VS becomes widely used in computer-based search for novel lead molecules. Typically, there are two approaches to the general problem: virtual screening by docking, when the 3D structure of the biological target (protein or enzyme) involved in the disease is available, and similarity-based virtual screening, where no information on the protein is necessary, instead, structural information of one or more known (bind to protein) molecules are used as structural query. The screening procedure retrieves molecules from the database according to the molecular similarity principle which states that structurally similar molecules exhibit similar biological activities. Similarity searches are now a standard tool for drug discovery. The idea behind such searches is that, given a compound with an interesting biological activity is compared to other compounds. The basic idea of similarity-based Virtual Screening is a very simple and it was first enunciated explicitly by Johnson and Maggiora [2]; in which Similar Property Principle states that molecules that are structurally similar are likely to have similar properties. The main goal of any system for similarity based screening is to quantify the degree of similarity or resemblance between reference structure (target query or queries) and each of the structures in database that is being screened for both real and virtual screening. A similarity measure requires three components: the molecules’ representation that is used to characterize them when are being compared, the weighting scheme that priorities the importance of various components of these representations and the coefficient that is used to calculate the degree of similarity or relatedness between two structural representations. This paper suggests a new ligand-based VS approach for similarity search. The new approach is based on the relationship between the target’s molecules features and all molecules’ features of in database. In the next section, fuzzy text retrieval method is explained. In section 3, we overviews some related works belong to this area. Materials and methods are discussed in section 4, including our proposed method FCC and all our experiments. In section 5, results were presented including evaluation of the new method based on measurement of recall of active molecules. Finally, our conclusions are presented in section 6.

Similarity searching

AbstractSimilarity searching is one of the traditional and most widely applied approaches in chemical and pharmaceutical research to select compounds with desired properties from databases. The computational efficiency of many (but not all) similarity search techniques has further increased their popularity as compound databases began to rapidly grow in size. Different methods have been developed for small molecule similarity searching. However, foundations and intrinsic limitations of similarity searching are often not well understood, although a number of similarity methods are rather simplistic. Regardless of methodological details, all similarity search approaches depend on how molecular similarity is evaluated and quantified. In its essence, molecular similarity is a subjective concept and much dependent on how we represent and view molecular structures. Moreover, trying to understand the relationship between molecular similarity, however assessed, and structure‐dependent properties including, first and foremost, biological activity continues to be a challenging problem. Consequently, although similarity searching usually provides a quantitative readout and a ranking of compounds relative to chosen reference molecules, predicting structure–activity relationships on the basis of calculated similarity values often involves subjective criteria and chemical intuition. Thus, similarity searching is still far from being a routine application in database mining. In this review, we first discuss important principles underlying similarity searching, describe its tasks, and introduce major categories of search methods. Then, we focus on molecular fingerprints, the design and application of which can be regarded as a paradigm for the similarity search field. © 2011 John Wiley & Sons, Ltd. WIREs Comput Mol Sci 2011 1 260‐282 DOI: 10.1002/wcms.23This article is categorized under: Computer and Information Science > Databases and Expert Systems

An efficient video similarity search algorithm

For the convenience of content-based video retrieval in large storage device, a new efficient video similarity search approach is proposed in this paper. To solve two challenging problems: similarity measurement and search method, a novel video feature computation of image characteristic code was proposed based on spatial-temporal distribution of video frame sequences. The video similarity is measured based on the calculation of the number of similar video components. For the scalable computing requirement, a new search method according to clustering index table was presented by index clustering. The experimental results in large database query tests show this method is efficient and effective for similar video search.

Multiple Pharmacophore Models Combined with Molecular Docking: A Reliable Way for Efficiently Identifying Novel PDE4 Inhibitors with High Structural Diversity

Multiple pharmacophore models were constructed based on the 18 crystal structures of phosphodiesterase 4 (PDE4) in complex with different inhibitors for discovering new potential PDE4 inhibitors. After validation of their efficiency in screening, 10 of the pharmacophore models were confirmed effective. Remarkably, the hits retrieved by these effective pharmacophore models were different, demonstrating that different pharmacophore models may have different performances in database screening. Therefore, all these models were employed to screen the compound database SPECS for finding potent leads with much structural diversity. Combining all the screened hits based on the 10 pharmacophore models, followed by molecular docking and bioassay, 4 of 53 tested compounds were found as active as rolipram (a well studied PDE4 inhibitor). More impressively, the four potent inhibitors with different chemical scaffolds were discovered by three different pharmacophore models separately, suggesting that a single pharmacophore model-based screening might not be efficient in thoroughly identifying potential hits from a compound database. This study also revealed that ligand-receptor complex structure-based pharmacophore is more efficient for identifying potent hits with great structural diversity in comparison with ligand-based pharmacophore and similarity search approaches. Therefore, multiple pharmacophore model-based virtual screenings should be used, if available, in combination with molecular docking for fully discovering hit compounds from compound databases.

PPR based similarity search for gas monitoring data

A Piecewise Polynomial Representation (PPR) based similarity search approach for time series data in coal well gas monitoring was proposed. For experiments, sample data were real gas monitoring data obtained from Yuhua Coal Mine, and evaluation criterions were information loss and mean search efficiency. Experimental results compared with the performance of Discrete Fourier Transform (DFT) based approach and Discrete Wavelet Transform (DWT) based approach show: in the condition of the same compress rate, values of information loss of PPR, DFT and DWT based approaches are very close to each other; Whereas, mean search efficiency of PPR based approach is 32% higher than that of DFT based approach, and 34% higher than that of DWT based approach respectively. Therefore, PPR based approach is more suitable for similarity search of gas monitoring data.

Similarity-based virtual screening using bayesian inference network

Many methods have been developed to capture the biological similarity between two compounds for use in drug discovery. A variety of similarity metrics have been introduced, the Tanimoto coefficient being the most prominent. Many of the approaches assume that molecular features or descriptors that do not relate to the biological activity carry the same weight as the important aspects in terms of biological similarity. Herein, a novel similarity searching approach using a Bayesian inference network is discussed. Similarity searching is regarded as an inference or evidential reasoning process in which the probability that a given compound has biological similarity with the query is estimated and used as evidence. Our experiments demonstrate that the similarity approach based on Bayesian inference networks is likely to outperform the Tanimoto similarity search and offer a promising alternative to existing similarity search approaches.

Similarity‐Based Virtual Screening with a Bayesian Inference Network

Many methods have been developed to capture the biological similarity between two compounds for use in drug discovery. A variety of similarity metrics have been introduced, the Tanimoto coefficient being the most prominent. Many of the approaches assume that molecular features or descriptors that do not relate to the biological activity carry the same weight as the important aspects in terms of biological similarity. Herein, a novel similarity searching approach using a Bayesian inference network is discussed. Similarity searching is regarded as an inference or evidential reasoning process in which the probability that a given compound has biological similarity with the query is estimated and used as evidence. Our experiments demonstrate that the similarity approach based on Bayesian inference networks is likely to outperform the Tanimoto similarity search and offer a promising alternative to existing similarity search approaches.

Annotation of 390 bovine miRNA genes by sequence similarity with other species

Source/description: MicroRNAs (miRNAs) are small noncoding RNA molecules that negatively regulate the expression of many protein-coding genes by affecting the degradation rate of the mRNAs or the correct translation of the gene products. miRNAs are normally encoded in pre-miRNA sequences with lengths of approximately 70 nt; these in turn are processed into mature sequences of about 22 nt. In the latest ENSEMBL release of the cow genome (http://www.ensembl.org/Bos_taurus), there are 524 identified miRNA genes, but only 117 are annotated in MIRBASE (http://microrna.sanger.ac.uk/) with unique identifiers. The remaining 407 miRNA genes are annotated as novel genes by the ENSEMBL annotation pipeline. We investigated this last group of sequences, which are predicted pre-miRNA-encoding genes for which corresponding bovine miRNAs have not yet been identified. A similarity search approach was used to identify homologous sequences among the miRNAs annotated in other species.

Parallel Guided Dynamic Programming Approach for DNA Sequence Similarity Search

Parallel Guided Dynamic Programming Approach for DNA Sequence Similarity Search

Approach for effective fractal-based similarity search of stochastic non-stationary time series

Traditional dimension reduction methods about similarity query introduce the smoothness to data series in some degree that the important features of time series about non-linearity and fractal are destroyed.A high-precision random non-stationary time series method named FSPA was proposed based on fractal theory and R/S analysis,which retained a non-linear time series and important fractal characteristics,and realized the reduction of the dimensions.The experiments have been performed on synthetic,as well as real data sequences to evaluate the proposed method,and the results indicate that the method has higher accuracy and requires less storage space.

Data Mining a Small Molecule Drug Screening Representative Subset from NIH PubChem

PubChem is a scientific showcase of the NIH Roadmap Initiatives. It is a compound repository created to facilitate information exchange and data sharing among the NIH Roadmap-funded Molecular Library Screening Center Network (MLSCN) and the scientific community. However, PubChem has more than 10 million records of compound information. It will be challenging to conduct a drug screening of the whole database of millions of compounds. Thus, the purpose of the present study was to develop a data mining cheminformatics approach in order to construct a representative and structure-diverse sublibrary from the large PubChem database. In this study, a new chemical diverse representative subset, rePubChem, was selected by whole-molecule chemistry-space matrix calculation using the cell-based partition algorithm. The representative subset was generated and was then subjected to evaluations by compound property analyses based on 1D and 2D molecular descriptors. The new subset was also examined and assessed for self-similarity analysis based on 2D molecular fingerprints in comparing with the source compound library. The new subset has a much smaller library size (540K compounds) with minimum similarity and redundancy without loss of the structural diversity and basic molecular properties of its parent library (5.3 million compounds). The new representative subset library generated could be a valuable structure-diverse compound resource for in silico virtual screening and in vitro HTS drug screening. In addition, the established subset generation method of using the combined cell-based chemistry-space partition metrics with pairwised 2D fingerprint-based similarity search approaches will also be important to a broad scientific community interested in acquiring structurally diverse compounds for efficient drug screening, building representative virtual combinatorial chemistry libraries for syntheses, and data mining large compound databases like the PubChem library in general.

Novel approaches for small biomolecule classification and structural similarity search

Structural similarity search among small molecules is a standard tool used in molecular classification and in-silico drug discovery. The effectiveness of this general approach depends on how well the following problems are addressed. The notion of similarity should be chosen for providing the highest level of discrimination of compounds with respect to the bioactivity of interest. The data structure for performing search should be very efficient as the molecular databases of interest include several millions of compounds. In this paper we summarize the recent applications of k -nearest-neighbor search method for small molecule classification. The k -nn classification of small molecules is based on selecting the most relevant set of chemical descriptors which are then compared under standard Minkowski distance L p . Here we describe how to computationally design the optimal weighted Minkowski distance wL p for maximizing the discrimination between active and inactive compounds wrt bioactivities of interest. k -nn classification requires fast similarity search for predicting bioactivity of a new molecule. We then focus on construction of pruning based k -nn search data structures for any wL p distance that minimizes similarity search time. The accuracy achieved by k -nn classifier is better than the alternative LDA and MLR approaches and is comparable to the ANN methods. In terms of running time, k -nn classifier is considerably faster than the ANN approach especially when large data sets are used. Furthermore, k -nn classifier is capable of quantification of the level of bioactivity rather than returning a binary decision and can bring more insight to the nature of the activity via eliminating unrelated descriptors of the compounds with respect to the activity in question.

Generation, annotation, analysis and database integration of 16,500 white spruce EST clusters

BackgroundThe sequencing and analysis of ESTs is for now the only practical approach for large-scale gene discovery and annotation in conifers because their very large genomes are unlikely to be sequenced in the near future. Our objective was to produce extensive collections of ESTs and cDNA clones to support manufacture of cDNA microarrays and gene discovery in white spruce (Picea glauca [Moench] Voss).ResultsWe produced 16 cDNA libraries from different tissues and a variety of treatments, and partially sequenced 50,000 cDNA clones. High quality 3' and 5' reads were assembled into 16,578 consensus sequences, 45% of which represented full length inserts. Consensus sequences derived from 5' and 3' reads of the same cDNA clone were linked to define 14,471 transcripts. A large proportion (84%) of the spruce sequences matched a pine sequence, but only 68% of the spruce transcripts had homologs in Arabidopsis or rice. Nearly all the sequences that matched the Populus trichocarpa genome (the only sequenced tree genome) also matched rice or Arabidopsis genomes. We used several sequence similarity search approaches for assignment of putative functions, including blast searches against general and specialized databases (transcription factors, cell wall related proteins), Gene Ontology term assignation and Hidden Markov Model searches against PFAM protein families and domains. In total, 70% of the spruce transcripts displayed matches to proteins of known or unknown function in the Uniref100 database (blastx e-value < 1e-10). We identified multigenic families that appeared larger in spruce than in the Arabidopsis or rice genomes. Detailed analysis of translationally controlled tumour proteins and S-adenosylmethionine synthetase families confirmed a twofold size difference. Sequences and annotations were organized in a dedicated database, SpruceDB. Several search tools were developed to mine the data either based on their occurrence in the cDNA libraries or on functional annotations.ConclusionThis report illustrates specific approaches for large-scale gene discovery and annotation in an organism that is very distantly related to any of the fully sequenced genomes. The ArboreaSet sequences and cDNA clones represent a valuable resource for investigations ranging from plant comparative genomics to applied conifer genetics.

VECTOR SPACE INDEXING FOR BIOSEQUENCE SIMILARITY SEARCHES

We present a multi-dimensional indexing approach for fast sequence similarity search in DNA and protein databases. In particular, we propose effective transformations of subsequences into numerical vector domains and build efficient index structures on the transformed vectors. We then define distance functions in the transformed domain and examine properties of these functions. We experimentally compared their (a) approximation quality for k-Nearest Neighbor (k-NN) queries and both (b) pruning ability and (c) approximation quality for ε-range queries. Results for k-NN queries, which we present here, show that our proposed distances FD2 and WD2 (i.e. Frequency and Wavelet Distance functions for 2-grams) perform significantly better than the others. We then develop effective index structures, based on R-trees and scalar quantization, on top of transformed vectors and distance functions. Promising results from the experiments on real biosequence data sets are presented.

Artificial Neural Networks Based Fingerprint Authentication With Clusters Algorithm

A novel and fast fingerprint identification technique is presented in this paper by using ANN. The proposed method use a novel clustering algorithm to detect similar feature groups from multiple template images generated from the same finger and create the cluster core set. The proposed feature extraction scheme is based on the reduction of information contents to the require minimum and define which part of the image is crucial and which is omitted. In the new method, a quick response was achieved by manipulating the search order inside the experimental databases. The novelty of our proposed method is to find the configuration of ANN system that could provide good generalization ability and sufficient discrimination capability at the same time. To achieve that goal, we have introduced new supervised recurrent neural-network. The experiments results demonstrate that this similarity search approach with ANN proves suitable one-to many matching of fingerprints on large databases. Povzetek: Hitro prepoznavanje prstnih odtisov je realizirano z ANN in grupiranjem.

A multi-step approach for partial similarity search in large image data using histogram intersection

We investigate the problem of retrieving partially similar images from a large image database. The region-based image retrieval technique is a method of retrieving partially similar images and has been proposed as a way to efficiently process queries in an image database In region-based image retrieval, region matching is indispensable for computing the partial similarity between two images because the query processing is based upon regions instead of the entire image. A naive method of region matching is a pairwise comparison between regions; this causes severe overhead and deteriorates the performance of query processing. In this paper, we focus on the development of a filtering function for the reduction of overall search time in region-based image retrieval, which is of special importance in the case of retrieving partially similar images from a large image database. To prune irrelevant images in a database, we introduce a correct and efficient similarity function by using the Histogram Intersection, which is needed for a crude selection based on a lower bounding property. Subsequently the result is refined by the pairwise region comparison between the query image and selected images. We have performed extensive experiments on synthetic and real image data to evaluate our proposed method. The experimental results reveal that our proposed technique achieves a significant pruning of up to 99% of irrelevant images and is up to 22 times faster than pairwise comparison, where the number of bins is set at 100.

A multistep approach for shape similarity search in image databases

Shape similarity searching is a crucial task in image databases, particularly in the presence of errors induced by segmentation or scanning images. The resulting slight displacements or rotations have not been considered so far in the literature. We present a new similarity model that flexibly addresses this problem. By specifying neighborhood influence weights, the user may adapt the similarity distance functions to his or her own requirements or preferences. Technically, the new similarity model is based on quadratic forms for which we present a multi-step query processing architecture, particularly for high dimensions as they occur in image databases. Our algorithm to reduce the dimensionality of quadratic form-based similarity queries results in a lower-bounding distance function that is proven to provide an optimal filter selectivity. Experiments on our test database of 10,000 images demonstrate the applicability and the performance of our approach, even in dimensions as high as 1,024.

Using extended feature objects for partial similarity retrieval

In this paper, we introduce the concept of extended feature objects for similarity retrieval. Conventional approaches for similarity search in databases map each object in the database to a point in some high-dimensional feature space and define similarity as some distance measure in this space. For many similarity search problems, this feature-based approach is not sufficient. When retrieving partially similar polygons, for example, the search cannot be restricted to edge sequences, since similar polygon sections may start and end anywhere on the edges of the polygons. In general, inherently continuous problems such as the partial similarity search cannot be solved by using point objects in feature space. In our solution, we therefore introduce extended feature objects consisting of an infinite set of feature points. For an efficient storage and retrieval of the extended feature objects, we determine the minimal bounding boxes of the feature objects in multidimensional space and store these boxes using a spatial access structure. In our concrete polygon problem, sets of polygon sections are mapped to 2D feature objects in high-dimensional space which are then approximated by minimal bounding boxes and stored in an R $^*$ -tree. The selectivity of the index is improved by using an adaptive decomposition of very large feature objects and a dynamic joining of small feature objects. For the polygon problem, translation, rotation, and scaling invariance is achieved by using the Fourier-transformed curvature of the normalized polygon sections. In contrast to vertex-based algorithms, our algorithm guarantees that no false dismissals may occur and additionally provides fast search times for realistic database sizes. We evaluate our method using real polygon data of a supplier for the car manufacturing industry.