Pol II promoter prediction using characteristic 4-mer motifs: a machine learning approach

Anwar, Firoz; Baker, Syed Murtuza; Jabid, Taskeed; Hasan, Md. Mehedi; Shoyaib, Mohammad; Khan, Haseena; Walshe, Ray

doi:10.1186/1471-2105-9-414

Cited by 38 publications

(29 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our 10-fold cross-validated accuracy is 92.1% for H. sapiens. Recently, by use of promoters and CDSs as benchmark data set, a SVM-based method was developed to discriminate promoter sequences from nonpromoter sequences (Anwar et al 2008 H. sapiens were achieved using 7-fold cross-validation. It must be noted that intron was not considered here for prediction.…”

Section: Comparison Accuraciesmentioning

confidence: 99%

“…Consequently, some on-line available tools, such as Eponine (Down and Hubbard 2002), Corepromoter (Zhang 2005), CpGProD (Ponger and Mouchiroud 2002), Promoter2.0 (Knudsen 1999), FirstEF (Davuluri et al 2001), promH , Promoter Scan (Prestridge 1995), Dragon promoter finder (Bajic et al 2002), NNPP2.2 (Reese 2001), McPromoter (Ohler 2006), ARTS (Sonnenburg et al 2006), ProSOM (Abeel et al 2008a, b), and RBF-TSS (Mahdi and Rouchka 2009) have been designed for the detection of promoters. Although contemporary methods have achieved great progress in promoter recognition, they were still limited in predictive performance (Anwar et al 2008;Abeel et al 2008a, b;Yang et al 2008). The performances of these methods are unreliable with poor specificity or poor sensitivity.…”

Section: Introductionmentioning

confidence: 98%

“…In the past 20 years, many algorithms, such as position weight matrix (PWM) (Prestridge 1995;Huerta and Collado-Vides 2003;Li and Lin 2006;Akan and Deloukas 2008), hidden Markov model (HMM) (Pedersen et al 1996;Ohler et al 1999Ohler et al , 2001Ohler et al , 2002Ohler 2006), artificial neural network (NN) (Horton and Kanehisa 1992;Pedersen and Engelbrecht 1995;Reese 2001;Burden et al 2005), and support vector machine (SVM) (Gordon et al 2003(Gordon et al , 2006Gangal and Sharma 2005;Anwar et al 2008) have been developed to identify promoters in eukaryotic and prokaryotic genomes using CpG islands, hexamer compositions, transcription factor binding sites (TFBS) and other information. Consequently, some on-line available tools, such as Eponine (Down and Hubbard 2002), Corepromoter (Zhang 2005), CpGProD (Ponger and Mouchiroud 2002), Promoter2.0 (Knudsen 1999), FirstEF (Davuluri et al 2001), promH , Promoter Scan (Prestridge 1995), Dragon promoter finder (Bajic et al 2002), NNPP2.2 (Reese 2001), McPromoter (Ohler 2006), ARTS (Sonnenburg et al 2006), ProSOM (Abeel et al 2008a, b), and RBF-TSS (Mahdi and Rouchka 2009) have been designed for the detection of promoters.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Eukaryotic and prokaryotic promoter prediction using hybrid approach

Lin

2010

Theory Biosci.

View full text Add to dashboard Cite

Promoters are modular DNA structures containing complex regulatory elements required for gene transcription initiation. Hence, the identification of promoters using machine learning approach is very important for improving genome annotation and understanding transcriptional regulation. In recent years, many methods have been proposed for the prediction of eukaryotic and prokaryotic promoters. However, the performances of these methods are still far from being satisfactory. In this article, we develop a hybrid approach (called IPMD) that combines position correlation score function and increment of diversity with modified Mahalanobis Discriminant to predict eukaryotic and prokaryotic promoters. By applying the proposed method to Drosophila melanogaster, Homo sapiens, Caenorhabditis elegans, Escherichia coli, and Bacillus subtilis promoter sequences, we achieve the sensitivities and specificities of 90.6 and 97.4% for D. melanogaster, 88.1 and 94.1% for H. sapiens, 83.3 and 95.2% for C. elegans, 84.9 and 91.4% for E. coli, as well as 80.4 and 91.3% for B. subtilis. The high accuracies indicate that the IPMD is an efficient method for the identification of eukaryotic and prokaryotic promoters. This approach can also be extended to predict other species promoters.

show abstract

Section: Comparison Accuraciesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Eukaryotic and prokaryotic promoter prediction using hybrid approach

Lin

2010

Theory Biosci.

View full text Add to dashboard Cite

show abstract

“…Initially, for all four nucleotides (A, T, C and G), different window lengths were generated [2]. In order to produce different features based on nucleotides, the frequency of occurrence of the four nucleotide bases was considered, and this method also created a standard window size among all the selected sequences.…”

Section: Featuresmentioning

confidence: 99%

Machine Learning Approaches for Prediction of Expansin Gene Family in Indica Rice

Hemalatha¹,

Rajesh

Narayanan

2013

Agric Res

View full text Add to dashboard Cite

Expansin refers to a family of proteins present in the plant cell wall which has important roles in plant cell growth, emergence of root hairs, meristem function and other developmental processes. A major constraint to rice production is submergence of rice by flash flooding. In our earlier study, we had identified 21 novel sequences related to expansin gene families in the genome of indica rice using genome-wide analysis. Development of a tool for the prediction of these expansin genes using computational approaches might significantly enhance rice gene annotation. ExpansinPred, a novel computational method based on radial basis function (RBF) and support vector machines (SVMs) for prediction of aexpansins (EXPA) and b-expansins (EXPB), is presented in this work. Two large families of expansin genes have been discovered in plants, namely EXPA and EXPB. The experimental data are curated from NCBI and include 24 EXPA and 20 EXPB, of indica rice, after redundancy elimination. The proper window length for a potential expansin was optimized as 4 for EXPA and EXPB with prediction accuracies 100 % each for both classifiers for RBF classifier. For SVM, the window length was optimized as 3 for EXPA and 4 for EXPB with prediction accuracies 90 and 100 %, respectively. To evaluate the prediction performance of ExpansinPred, cross-validation, independent dataset validation and jackknife validation were carried out. ExpansinPred was also compared with four more algorithms namely Naïve Bayes, sequential minimal optimization, J48 and random forest. To further prove that species-specific predictor is much better than general tool, ExpansinPred was compared with an All-plant tool and also with plants other than rice as test set. The different statistical analyses carried out demonstrated that the proposed algorithm is a useful computational tool for rice genome annotation, specifically for predicting expansin gene family, and can benefit rice research community.

show abstract

“…In the field of life sciences, SVM is also a powerful tool to build effective predicting models in bioinformatics and computational systems biology, such as protein structure and stability prediction [13,14], RNA secondary structure prediction [15], bacterial transcription start sites prediction [16], virtual screening for drug discovering [17][18][19][20], drug metabolism prediction [21], disease prognosis and prediction [22,23], as well as promoter recognition and structure analysis [24][25][26][27][28][29][30][31][32]. However, SVM has not been reported to use in predicting the strength of promoter or even the regulatory elements.…”

Section: Introductionmentioning

confidence: 99%

Construction of precise support vector machine based models for predicting promoter strength

Meng

Mai

et al. 2017

Quant. Biol.

View full text Add to dashboard Cite

Background: The prediction of the prokaryotic promoter strength based on its sequence is of great importance not only in the fundamental research of life sciences but also in the applied aspect of synthetic biology. Much advance has been made to build quantitative models for strength prediction, especially the introduction of machine learning methods such as artificial neural network (ANN) has significantly improve the prediction accuracy. As one of the most important machine learning methods, support vector machine (SVM) is more powerful to learn knowledge from small sample dataset and thus supposed to work in this problem. Methods: To confirm this, we constructed SVM based models to quantitatively predict the promoter strength. A library of 100 promoter sequences and strength values was randomly divided into two datasets, including a training set (≥10 sequences) for model training and a test set (≥10 sequences) for model test. Results:The results indicate that the prediction performance increases with an increase of the size of training set, and the best performance was achieved at the size of 90 sequences. After optimization of the model parameters, a highperformance model was finally trained, with a high squared correlation coefficient for fitting the training set (R 2 > 0.99) and the test set (R 2 > 0.98), both of which are better than that of ANN obtained by our previous work.Conclusions: Our results demonstrate the SVM-based models can be employed for the quantitative prediction of promoter strength.

show abstract

Pol II promoter prediction using characteristic 4-mer motifs: a machine learning approach

Cited by 38 publications

References 26 publications

Eukaryotic and prokaryotic promoter prediction using hybrid approach

Eukaryotic and prokaryotic promoter prediction using hybrid approach

Machine Learning Approaches for Prediction of Expansin Gene Family in Indica Rice

Construction of precise support vector machine based models for predicting promoter strength

Contact Info

Product

Resources

About