Search

This chapter is based on exploiting the network-based representations of proteins, metagraphs, in protein-protein interaction network to identify candidate disease-causing proteins. Protein-protein interaction (PPI) networks are effective tools in studying the functional roles of proteins in the development of various diseases. However, they are insufficient without the support of additional biological knowledge for proteins such as their molecular functions and biological processes. To enhance PPI networks, we utilize biological properties of individual proteins as well. More specifically, we integrate keywords from UniProt database describing protein properties into the PPI network and construct a novel heterogeneous PPI-Keyword (PPIK) network consisting of both proteins and keywords. As proteins with similar functional duties or involving in the same metabolic pathway tend to have similar topological characteristics, we propose to represent them with metagraphs. Compared to the traditional network motif or subgraph, a metagraph can capture the topological arrangements through not only the protein-protein interactions but also protein-keyword associations. We feed those novel metagraph representations into classifiers for disease protein prediction and conduct our experiments on three different PPI databases. They show that the proposed method consistently increases disease protein prediction performance across various classifiers, by 15.3% in AUC on average. It outperforms the diffusion-based (e.g., RWR) and the module-based baselines by 13.8-32.9% in overall disease protein prediction. Breast cancer protein prediction outperforms RWR, PRINCE, and the module-based baselines by 6.6-14.2%. Finally, our predictions also exhibit better correlations with literature findings from PubMed database.

Experimental prediction of drug-target interactions is expensive, time-consuming and tedious. Fortunately, computational methods help narrow down the search space for interaction candidates to be further examined via wet-lab techniques. Nowadays, the number of attributes/features for drugs and targets, as well as the amount of their interactions, are increasing, making these computational methods inefficient or occasionally prohibitive. This motivates us to derive a reduced feature set for prediction. In addition, since ensemble learning techniques are widely used to improve the classification performance, it is also worthwhile to design an ensemble learning framework to enhance the performance for drug-target interaction prediction. In this paper, we propose a framework for drug-target interaction prediction leveraging both feature dimensionality reduction and ensemble learning. First, we conducted feature subspacing to inject diversity into the classifier ensemble. Second, we applied three different dimensionality reduction methods to the subspaced features. Third, we trained homogeneous base learners with the reduced features and then aggregated their scores to derive the final predictions. For base learners, we selected two classifiers, namely Decision Tree and Kernel Ridge Regression, resulting in two variants of ensemble models, EnsemDT and EnsemKRR, respectively. In our experiments, we utilized AUC (Area under ROC Curve) as an evaluation metric. We compared our proposed methods with various state-of-the-art methods under 5-fold cross validation. Experimental results showed EnsemKRR achieving the highest AUC (94.3%) for predicting drug-target interactions. In addition, dimensionality reduction helped improve the performance of EnsemDT. In conclusion, our proposed methods produced significant improvements for drug-target interaction prediction., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Experimental determination of drug-target interactions is expensive and time-consuming. Therefore, there is a continuous demand for more accurate predictions of interactions using computational techniques. Algorithms have been devised to infer novel interactions on a global scale where the input to these algorithms is a drug-target network (i.e., a bipartite graph where edges connect pairs of drugs and targets that are known to interact). However, these algorithms had difficulty predicting interactions involving new drugs or targets for which there are no known interactions (i.e., "orphan" nodes in the network). Since data usually lie on or near to low-dimensional non-linear manifolds, we propose two matrix factorization methods that use graph regularization in order to learn such manifolds. In addition, considering that many of the non-occurring edges in the network are actually unknown or missing cases, we developed a preprocessing step to enhance predictions in the "new drug" and "new target" cases by adding edges with intermediate interaction likelihood scores. In our cross validation experiments, our methods achieved better results than three other state-of-the-art methods in most cases. Finally, we simulated some "new drug" and "new target" cases and found that GRMF predicted the left-out interactions reasonably well.

Background: Multiple computational methods for predicting drug-target interactions have been developed to facilitate the drug discovery process. These methods use available data on known drug-target interactions to train classifiers with the purpose of predicting new undiscovered interactions. However, a key challenge regarding this data that has not yet been addressed by these methods, namely class imbalance, is potentially degrading the prediction performance. Class imbalance can be divided into two sub-problems. Firstly, the number of known interacting drug-target pairs is much smaller than that of non-interacting drug-target pairs. This imbalance ratio between interacting and non-interacting drug-target pairs is referred to as the between-class imbalance. Between-class imbalance degrades prediction performance due to the bias in prediction results towards the majority class (i.e. the non-interacting pairs), leading to more prediction errors in the minority class (i.e. the interacting pairs). Secondly, there are multiple types of drug-target interactions in the data with some types having relatively fewer members (or are less represented) than others. This variation in representation of the different interaction types leads to another kind of imbalance referred to as the within-class imbalance. In within-class imbalance, prediction results are biased towards the better represented interaction types, leading to more prediction errors in the less represented interaction types., Results: We propose an ensemble learning method that incorporates techniques to address the issues of between-class imbalance and within-class imbalance. Experiments show that the proposed method improves results over 4 state-of-the-art methods. In addition, we simulated cases for new drugs and targets to see how our method would perform in predicting their interactions. New drugs and targets are those for which no prior interactions are known. Our method displayed satisfactory prediction performance and was able to predict many of the interactions successfully., Conclusions: Our proposed method has improved the prediction performance over the existing work, thus proving the importance of addressing problems pertaining to class imbalance in the data.

While high-throughput technologies are expected to play a critical role in clinical translational research for complex disease diagnosis, the ability to accurately and consistently discriminate disease phenotypes by determining the gene and protein expression patterns as signatures of different clinical conditions remains a challenge in translational bioinformatics. In this study, we propose a novel feature selection algorithm: Multi-Resolution-Test (MRT-test) that can produce significantly accurate and consistent phenotype discrimination across a series of omics data. Our algorithm can capture those features contributing to subtle data behaviors instead of selecting the features contributing to global data behaviors, which seems to be essential in achieving clinical level diagnosis for different expression data. Furthermore, as an effective biomarker discovery algorithm, it can achieve linear separation for high-dimensional omics data with few biomarkers. We apply our MRT-test to complex disease phenotype diagnosis by combining it with state-of-the-art classifiers and attain exceptional diagnostic results, which suggests that our method's advantage in molecular diagnostics. Experimental evaluation showed that MRT-test based diagnosis is able to generate consistent and robust clinical-level phenotype separation for various diseases. In addition, based on the seed biomarkers detected by the MRT-test, we design a novel network marker synthesis (NMS) algorithm to decipher the underlying molecular mechanisms of tumorigenesis from a systems viewpoint. Unlike existing top-down gene network building approaches, our network marker synthesis method has a less dependence on the global network and enables it to capture the gene regulators for different subnetwork markers, which will provide biologically meaningful insights for understanding the genetic basis of complex diseases.

Motivation: Protein complexes are of great importance for unraveling the secrets of cellular organization and function. The AP-MS technique has provided an effective high-throughput screening to directly measure the co-complex relationship among multiple proteins, but its performance suffers from both false positives and false negatives. To computationally predict complexes from AP-MS data, most existing approaches either required the additional knowledge from known complexes (supervised learning), or had numerous parameters to tune., Method: In this article, we propose a novel unsupervised approach, without relying on the knowledge of existing complexes. Our method probabilistically calculates the affinity between two proteins, where the affinity score is evaluated by a co-complexed score or C2S in brief. In particular, our method measures the log-likelihood ratio of two proteins being co-complexed to being drawn randomly, and we then predict protein complexes by applying hierarchical clustering algorithm on the C2S score matrix., Results: Compared with existing approaches, our approach is computationally efficient and easy to implement. It has just one parameter to set and its value has little effect on the results. It can be applied to different species as long as the AP-MS data are available. Despite its simplicity, it is competitive or superior in performance over many aspects when compared with the state-of-the-art predictions performed by supervised or unsupervised approaches.

Background: Although high-throughput microarray based molecular diagnostic technologies show a great promise in cancer diagnosis, it is still far from a clinical application due to its low and instable sensitivities and specificities in cancer molecular pattern recognition. In fact, high-dimensional and heterogeneous tumor profiles challenge current machine learning methodologies for its small number of samples and large or even huge number of variables (genes). This naturally calls for the use of an effective feature selection in microarray data classification., Methods: We propose a novel feature selection method: multi-resolution independent component analysis (MICA) for large-scale gene expression data. This method overcomes the weak points of the widely used transform-based feature selection methods such as principal component analysis (PCA), independent component analysis (ICA), and nonnegative matrix factorization (NMF) by avoiding their global feature-selection mechanism. In addition to demonstrating the effectiveness of the multi-resolution independent component analysis in meaningful biomarker discovery, we present a multi-resolution independent component analysis based support vector machines (MICA-SVM) and linear discriminant analysis (MICA-LDA) to attain high-performance classifications in low-dimensional spaces., Results: We have demonstrated the superiority and stability of our algorithms by performing comprehensive experimental comparisons with nine state-of-the-art algorithms on six high-dimensional heterogeneous profiles under cross validations. Our classification algorithms, especially, MICA-SVM, not only accomplish clinical or near-clinical level sensitivities and specificities, but also show strong performance stability over its peers in classification. Software that implements the major algorithm and data sets on which this paper focuses are freely available at https://sites.google.com/site/heyaumapbc2011/., Conclusions: This work suggests a new direction to accelerate microarray technologies into a clinical routine through building a high-performance classifier to attain clinical-level sensitivities and specificities by treating an input profile as a 'profile-biomarker'. The multi-resolution data analysis based redundant global feature suppressing and effective local feature extraction also have a positive impact on large scale 'omics' data mining.

Multiprotein complexes play central roles in many cellular pathways. Although many high-throughput experimental techniques have already enabled systematic screening of pairwise protein-protein interactions en masse, the amount of experimentally determined protein complex data has remained relatively lacking. As such, researchers have begun to exploit the vast amount of pairwise interaction data to help discover new protein complexes. However, mining for protein complexes in interaction networks is not an easy task because there are many data artefacts in the underlying protein-protein interaction data due to the limitations in the current high-throughput screening methods. We propose a novel DECAFF (Dense-neighborhood Extraction using Connectivity and conFidence Features) algorithm to mine for dense and reliable subgraphs in protein interaction networks. Our method is devised to address two major limitations in current high throughout protein interaction data, namely, incompleteness and high data noise. Experimental results with yeast protein interaction data show that the interaction subgraphs discovered by DECAFF matched significantly better with actual protein complexes than other existing approaches. Our results demonstrate that pairwise protein interaction networks can be effectively mined to discover new protein complexes, provided that the data artefacts in the underlying interaction data are taken into account adequately.

While recent technological advances have made available large datasets of experimentally-detected pairwise protein-protein interactions, there is still a lack of experimentally-determined protein complex data. To make up for this lack of protein complex data, we explore the mining of existing protein interaction graphs for protein complexes. This paper proposes a novel graph mining algorithm to detect the dense neighborhoods (highly connected regions) in an interaction graph which may correspond to protein complexes. Our algorithm first locates local cliques for each graph vertex (protein) and then merge the detected local cliques according to their affinity to form maximal dense regions. We present experimental results with yeast protein interaction data to demonstrate the effectiveness of our proposed method. Compared with other existing techniques, our predicted complexes can match or overlap significantly better with the known protein complexes in the MIPS benchmark database. Novel protein complexes were also predicted to help biologists in their search for new protein complexes.

As an effective tool for data analysis, formal concept analysis (FCA) is widely used in software engineering and machine learning. The construction of concept lattice is a key step of the FCA. How to effectively to update the concept lattice is still an open, interesting and important issue. To resolve this problem, an incremental algorithm for concept lattice on image structure similarity (SsimAddExten) was presented. The proposed method mapped each knowledge class on the conceptlattice into a graphic, when a new object was added or deleted in a knowledge class, the boundary profile of graphic will be changed, the graphic edge structure similarity was introduced as the calculation index of the change degree before and after the knowledge, and the concept lattice will be updated on the basis of the index. We performed experiments to test SsimAddExtent, whose computational efficiency obtains obvious advantages over mainstream methods on almost all test points, especially on the data set with a large number of attributes. But, its complexity is not reduced compared with mainstream methods. Both theoretical analysis and performance test show SsimAddExtent algorithm is better choice when we apply the FCA to large scale data or non-sparse data. [ABSTRACT FROM AUTHOR]

Data in the form of graphs are prevalent, ranging from biological and social networks to citation graphs and the Web. In particular, most real-world graphs are heterogeneous, containing objects of multiple types, which present new opportunities for many problems on graphs. Consider a typical proximity search problem on graphs, which boils down to measuring the proximity between two given nodes. Most earlier studies on homogeneous or bipartite graphs only measure a generic form of proximity, without accounting for different “semantic classes”—for instance, on a social network two users can be close for different reasons, such as being classmates or family members, which represent two distinct semantic classes. Learning these semantic classes are made possible on heterogeneous graphs through the concept of metagraphs. In this study, we identify metagraphs as a novel and effective means to characterize the common structures for a desired class of proximity. Subsequently, we propose a family of metagraph-based proximity, and employ a learning-to-rank technique that automatically learns the right parameters to suit the desired semantic class. In terms of efficiency, we develop a symmetry-based matching algorithm to speed up the computation of metagraph instances. Empirically, extensive experiments reveal that our metagraph-based proximity substantially outperforms the best competitor by more than 10 percent, and our matching algorithm can reduce matching time by more than half. As a further generalization, we aim to derive a general node and edge representation for heterogeneous graphs, in order to support arbitrary machine learning tasks beyond proximity search. In particular, we propose the finer-grained anchored metagraph, which is capable of discriminating the roles of nodes within the same metagraph. Finally, further experiments on the general representation show that we can outperform the state of the art significantly and consistently across various machine learning tasks. [ABSTRACT FROM AUTHOR]

High frequency oscillations (HFOs) are considered as biomarker for epileptogenicity. Reliable automation of HFOs detection is necessary for rapid and objective analysis, and is determined by accurate computation of the baseline. Although most existing automated detectors measure baseline accurately in channels with rare HFOs, they lose accuracy in channels with frequent HFOs. Here, we proposed a novel algorithm using the maximum distributed peak points method to improve baseline determination accuracy in channels with wide HFOs activity ranges and calculate a dynamic baseline. Interictal ripples (80-200Hz), fast ripples (FRs, 200-500Hz) and baselines in intracerebral EEGs from seven patients with intractable epilepsy were identified by experienced reviewers and by our computer-automated program, and the results were compared. We also compared the performance of our detector to four well-known detectors integrated in RIPPLELAB. The sensitivity and specificity of our detector were, respectively, 71% and 75% for ripples and 66% and 84% for FRs. Spearman's rank correlation coefficient comparing automated and manual detection was for ripples and for FRs (). In comparison to other detectors, our detector had a relatively higher sensitivity and specificity. In conclusion, our automated detector is able to accurately calculate a dynamic iEEG baseline in different HFO activity channels using the maximum distributed peak points method, resulting in higher sensitivity and specificity than other available HFO detectors. [ABSTRACT FROM AUTHOR]

The identification of network motifs has important applications in numerous domains, such as pattern detection in biological networks and graph analysis in digital circuits. However, mining network motifs is computationally challenging, as it requires enumerating subgraphs from a real-life graph, and computing the frequency of each subgraph in a large number of random graphs. In particular, existing solutions often require days to derive network motifs from biological networks with only a few thousand vertices. To address this problem, this paper presents a novel study on network motif discovery using Graphical Processing Units (GPUs). The basic idea is to employ GPUs to parallelize a large number of subgraph matching tasks in computing subgraph frequencies from random graphs, so as to reduce the overall computation time of network motif discovery. We explore the design space of GPU-based subgraph matching algorithms, with careful analysis of several crucial factors (such as branch divergences and memory coalescing) that affect the performance of GPU programs. Based on our analysis, we develop a GPU-based solution that (i) considerably differs from existing CPU-based methods in how it enumerates subgraphs, and (ii) exploits the strengths of GPUs in terms of parallelism while mitigating their limitations in terms of the computation power per GPU core. With extensive experiments on a variety of biological networks, we show that our solution is up to two orders of magnitude faster than the best CPU-based approach, and is around $20$ times more cost-effective than the latter, when taking into account the monetary costs of the CPU and GPUs used. [ABSTRACT FROM PUBLISHER]

Background: Identifying disease genes from human genome is an important but challenging task in biomedical research. Machine learning methods can be applied to discover new disease genes based on the known ones. Existing machine learning methods typically use the known disease genes as the positive training set P and the unknown genes as the negative training set N (non-disease gene set does not exist) to build classifiers to identify new disease genes from the unknown genes. However, such kind of classifiers is actually built from a noisy negative set N as there can be unknown disease genes in N itself. As a result, the classifiers do not perform as well as they could be.Result: Instead of treating the unknown genes as negative examples in N, we treat them as an unlabeled set U. We design a novel positive-unlabeled (PU) learning algorithm PUDI (PU learning for disease gene identification) to build a classifier using P and U. We first partition U into four sets, namely, reliable negative set RN, likely positive set LP, likely negative set LN and weak negative set WN. The weighted support vector machines are then used to build a multi-level classifier based on the four training sets and positive training set P to identify disease genes. Our experimental results demonstrate that our proposed PUDI algorithm outperformed the existing methods significantly.Conclusion: The proposed PUDI algorithm is able to identify disease genes more accurately by treating the unknown data more appropriately as unlabeled set U instead of negative set N. Given that many machine learning problems in biomedical research do involve positive and unlabeled data instead of negative data, it is possible that the machine learning methods for these problems can be further improved by adopting PU learning methods, as we have done here for disease gene identification.Availability and implementation: The executable program and data are available at http://www1.i2r.a-star.edu.sg/∼xlli/PUDI/PUDI.html.Contact: xlli@i2r.a-star.edu.sg or yang0293@e.ntu.edu.sgSupplementary information: Supplementary Data are available at Bioinformatics online. [ABSTRACT FROM AUTHOR]

Many cellular functions involve protein complexes that are formed by multiple interacting proteins. Tandem Affinity Purification (TAP) is a popular experimental method for detecting such multi-protein interactions. However, current computational methods that predict protein complexes from TAP data require converting the co-complex relationships in TAP data into binary interactions. The resulting pairwise protein-protein interaction (PPI) network is then mined for densely connected regions that are identified as putative protein complexes. Converting the TAP data into PPI data not only introduces errors but also loses useful information about the underlying multi-protein relationships that can be exploited to detect the internal organization (i.e., core-attachment structures) of protein complexes. In this article, we propose a method called CACHET that detects protein complexes with ore-ttament structures directly from bipartitAP data. CACHET models the TAP data as a bipartite graph in which the two vertex sets are the baits and the preys, respectively. The edges between the two vertex sets represent bait-prey relationships. CACHET first focuses on detecting high-quality protein-complex cores from the bipartite graph. To minimize the effects of false positive interactions, the bait-prey relationships are indexed with reliability scores. Only non-redundant, reliable bicliques computed from the TAP bipartite graph are regarded as protein-complex cores. CACHET constructs protein complexes by including attachment proteins into the cores. We applied CACHET on large-scale TAP datasets and found that CACHET outperformed existing methods in terms of prediction accuracy (i.e., F-measure and functional homogeneity of predicted complexes). In addition, the protein complexes predicted by CACHET are equipped with core-attachment structures that provide useful biological insights into the inherent functional organization of protein complexes. Our can be found at ; binary executables can also be found there. is also available at . [ABSTRACT FROM AUTHOR]

While in the process of decomposition of the Mallat discrete wavelet transform (DWT) fast algorithm, the algorithm has a drawback, that is, there is frequency distortion in the high frequency subband. In this paper, a new algorithm of decomposition and reconstruction in the discrete wavelet is presented. The algorithm can solve the frequency distortion in the high frequency subband in the process of decimation in each level. The simulation of numerical value example tests the validity of the algorithm. [ABSTRACT FROM AUTHOR]