The race for the discovery of enhancers at a genome-wide scale has been on since the commencement of next generation sequencing decades after the discovery of the first enhancer, SV40. A few enhancer-predicting features such as chromatin feature, histone modifications and sequence feature had been implemented with varying success rates. However, to date, there is no consensus yet on the single enhancer marker that can be employed to ultimately distinguish and uncover enhancers from the enormous genomic regions. Many supervised, unsupervised and semi-supervised computational approaches had emerged to complement and facilitate experimental approaches in enhancer discovery. In this review, we placed our focus on the recently emerged enhancer predictor tools that work on general enhancer features such as sequences, chromatin states and histone modifications, eRNA and of multiple feature approach. Comparisons of their prediction methods and outcomes were done across their functionally similar counterparts. We provide some recommendations and insights for future development of more comprehensive and robust tools.

BackgroundDiscrimination of transcription factor binding sites (TFBS) from background sequences plays a key role in computational motif discovery. Current clustering based algorithms employ homogeneous model for problem solving, which assumes that motifs and background signals can be equivalently characterized. This assumption has some limitations because both sequence signals have distinct properties.ResultsThis paper aims to develop a Self-Organizing Map (SOM) based clustering algorithm for extracting binding sites in DNA sequences. Our framework is based on a novel intra-node soft competitive procedure to achieve maximum discrimination of motifs from background signals in datasets. The intra-node competition is based on an adaptive weighting technique on two different signal models to better represent these two classes of signals. Using several real and artificial datasets, we compared our proposed method with several motif discovery tools. Compared to SOMBRERO, a state-of-the-art SOM based motif discovery tool, it is found that our algorithm can achieve significant improvements in the average precision rates (i.e., about 27%) on the real datasets without compromising its sensitivity. Our method also performed favourably comparing against other motif discovery tools.ConclusionsMotif discovery with model based clustering framework should consider the use of heterogeneous model to represent the two classes of signals in DNA sequences. Such heterogeneous model can achieve better signal discrimination compared to the homogeneous model.

Abstract-Convolutionary neural network (CNN) is a popular choice for supervised DNA motif prediction due to its excellent performances. To employ CNN, the input DNA sequences are required to be encoded as numerical values and represented as either vectors or multi-dimensional matrices. This paper evaluates a simple and more compact ordinal encoding method versus the popular one-hot encoding for DNA sequences. We compare the performances of both encoding methods using three sets of datasets enriched with DNA motifs. We found that the ordinal encoding performs comparable to the one-hot method but with significant reduction in training time. In addition, the one-hot encoding performances are rather consistent across various datasets but would require suitable CNN configuration to perform well. The ordinal encoding with matrix representation performs best in some of the evaluated datasets. This study implies that the performances of CNN for DNA motif discovery depends on the suitable design of the sequence encoding and representation. The good performances of the ordinal encoding method demonstrates that there are still rooms for improvement for the one-hot encoding method.

No abstract

Abstract.To detect or discover motifs in DNA sequences, two important concepts related to existing computational approaches are motif model and similarity score. One of motif models, represented by a position frequency matrix (PFM), has been widely employed to search for putative motifs. Detection and discovery of motifs can be done by comparing kmers with a motif model, or clustering kmers according to some criteria. In the past, information content based similarity scores have been widely used in searching tools. In this paper, we present a mismatchbased matrix similarity score (namely, MISCORE) for motif searching and discovering purpose. The proposed MISCORE can be biologically interpreted as an evolutionary metric for predicting a kmer as a motif member or not. Weighting factors, which are meaningful for biological data mining practice, are introduced in the MISCORE. The effectiveness of the MISCORE is investigated through exploring its separability, recognizability and robustness. Three well-known information contentbased matrix similarity scores are compared, and results show that our MISCORE works well.

Abstract-Convolutionary neural network (CNN) is a popular choice for supervised DNA motif prediction due to its excellent performances. To employ CNN, the input DNA sequences are required to be encoded as numerical values and represented as either vectors or multi-dimensional matrices. This paper evaluates a simple and more compact ordinal encoding method versus the popular one-hot encoding for DNA sequences. We compare the performances of both encoding methods using three sets of datasets enriched with DNA motifs. We found that the ordinal encoding performs comparable to the one-hot method but with significant reduction in training time. In addition, the one-hot encoding performances are rather consistent across various datasets but would require suitable CNN configuration to perform well. The ordinal encoding with matrix representation performs best in some of the evaluated datasets. This study implies that the performances of CNN for DNA motif discovery depends on the suitable design of the sequence encoding and representation. The good performances of the ordinal encoding method demonstrates that there are still rooms for improvement for the one-hot encoding method.

We propose an improved solution to the three-stage DNA motif prediction approach. The threestage approach uses only a subset of input sequences for initial motif prediction, and the initial motifs obtained are employed for site detection in the remaining input subset of non-overlaps. The currently available solution is not robust because motifs obtained from the initial subset are represented as a position weight matrices, which results in high false positives. Our approach, called DeepFinder, employs deep learning neural networks with features associated with binding sites to construct a motif model. Furthermore, multiple prediction tools are used in the initial motif prediction process to obtain a higher number of positive hits. Our features are engineered from the context of binding sites, which are assumed to be enriched with specificity information of sites recognized by transcription factor proteins. DeepFinder is evaluated using several performance metrics on ten chromatin immunoprecipitation (ChIP) datasets. The results show marked improvement of our solution in comparison with the existing solution. This indicates the effectiveness and potential of our proposed DeepFinder for large-scale motif analysis.

Discovery of motifs plays a key role in understanding