Prediction of DNA-binding proteins from relational features

Szabóová, Andrea; Kuželka, Ondřej; Železný, Filip; Tolar, Jakub

doi:10.1186/1477-5956-10-66

Cited by 7 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To obtain a reliable result with low mean square error, k -fold cross validation was always used in empirical works [7, 8]. In this study, 5-fold cross validation method was used to access the performance of each classifier on the main dataset.…”

Section: Methodsmentioning

confidence: 99%

Identification of DNA-Binding Proteins Using Support Vector Machine with Sequence Information

Xue

2013

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

DNA-binding proteins are fundamentally important in understanding cellular processes. Thus, the identification of DNA-binding proteins has the particularly important practical application in various fields, such as drug design. We have proposed a novel approach method for predicting DNA-binding proteins using only sequence information. The prediction model developed in this study is constructed by support vector machine-sequential minimal optimization (SVM-SMO) algorithm in conjunction with a hybrid feature. The hybrid feature is incorporating evolutionary information feature, physicochemical property feature, and two novel attributes. These two attributes use DNA-binding residues and nonbinding residues in a query protein to obtain DNA-binding propensity and nonbinding propensity. The results demonstrate that our SVM-SMO model achieves 0.67 Matthew's correlation coefficient (MCC) and 89.6% overall accuracy with 88.4% sensitivity and 90.8% specificity, respectively. Performance comparisons on various features indicate that two novel attributes contribute to the performance improvement. In addition, our SVM-SMO model achieves the best performance than state-of-the-art methods on independent test dataset.

show abstract

Section: Methodsmentioning

confidence: 99%

Identification of DNA-Binding Proteins Using Support Vector Machine with Sequence Information

Xue

2013

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

show abstract

“…A powerful feature of ILP is that, in addition to prediction, it automatically learns rules which can be readily understood. It has been successfully applied to predict and model various medical [ 14 , 15 ] and biological datasets [ 16 , 17 ]. However, the complexity and size of the hypothesis space often presents computational challenges in search time which limit both the insight and the predictive power of the rules found.…”

Section: Introductionmentioning

confidence: 99%

Automated identification of protein-ligand interaction features using Inductive Logic Programming: a hexose binding case study

et al. 2012

View full text Add to dashboard Cite

BackgroundThere is a need for automated methods to learn general features of the interactions of a ligand class with its diverse set of protein receptors. An appropriate machine learning approach is Inductive Logic Programming (ILP), which automatically generates comprehensible rules in addition to prediction. The development of ILP systems which can learn rules of the complexity required for studies on protein structure remains a challenge. In this work we use a new ILP system, ProGolem, and demonstrate its performance on learning features of hexose-protein interactions.ResultsThe rules induced by ProGolem detect interactions mediated by aromatics and by planar-polar residues, in addition to less common features such as the aromatic sandwich. The rules also reveal a previously unreported dependency for residues cys and leu. They also specify interactions involving aromatic and hydrogen bonding residues. This paper shows that Inductive Logic Programming implemented in ProGolem can derive rules giving structural features of protein/ligand interactions. Several of these rules are consistent with descriptions in the literature.ConclusionsIn addition to confirming literature results, ProGolem’s model has a 10-fold cross-validated predictive accuracy that is superior, at the 95% confidence level, to another ILP system previously used to study protein/hexose interactions and is comparable with state-of-the-art statistical learners.

show abstract