nDNA-prot: identification of DNA-binding proteins based on unbalanced classification

Abstract

BackgroundDNA-binding proteins are vital for the study of cellular processes. In recent genome engineering studies, the identification of proteins with certain functions has become increasingly important and needs to be performed rapidly and efficiently. In previous years, several approaches have been developed to improve the identification of DNA-binding proteins. However, the currently available resources are insufficient to accurately identify these proteins. Because of this, the previous research has been limited by the relatively unbalanced accuracy rate and the low identification success of the current methods.ResultsIn this paper, we explored the practicality of modelling DNA binding identification and simultaneously employed an ensemble classifier, and a new predictor (nDNA-Prot) was designed. The presented framework is comprised of two stages: a 188-dimension feature extraction method to obtain the protein structure and an ensemble classifier designated as imDC. Experiments using different datasets showed that our method is more successful than the traditional methods in identifying DNA-binding proteins. The identification was conducted using a feature that selected the minimum Redundancy and Maximum Relevance (mRMR). An accuracy rate of 95.80% and an Area Under the Curve (AUC) value of 0.986 were obtained in a cross validation. A test dataset was tested in our method and resulted in an 86% accuracy, versus a 76% using iDNA-Prot and a 68% accuracy using DNA-Prot.ConclusionsOur method can help to accurately identify DNA-binding proteins, and the web server is accessible at http://datamining.xmu.edu.cn/~songli/nDNA. In addition, we also predicted possible DNA-binding protein sequences in all of the sequences from the UniProtKB/Swiss-Prot database.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2105-15-298) contains supplementary material, which is available to authorized users.

imDC: an ensemble learning method for imbalanced classification with miRNA data

“…Since AdaBoost treats misclassified positive and negative examples equally, it cannot support high performance in imbalanced data. Asymboost is an improved algorithm of AdaBoost in which positive samples had a higher price when they were misclassified (Lin et al, 2013(Lin et al, , 2014Song et al, 2014). However, when positive and negative examples had an equal cost for classification, the algorithm was equivalent to AdaBoost.…”

Local-DPP: An improved DNA-binding protein prediction method by exploring local evolutionary information

“…Recently, Song et al [39] reported that the number of non-DNA-binding proteins in datasets far outweighs the number of DNA-binding proteins. They addressed the data imbalance problem by a novel ensemble classifier (imDC; see [39]).…”

“…To further improve DNA-binding protein prediction from the PseAAC vector, they also combined the PseAAC with a physicochemical distance transformation [27]. Besides PseAAC, DNA-binding proteins are represented by other commonly used sequence-based features, such as physicochemical properties [7,27,39,47], amino acid composition [7,42,49], autocross-covariance transformation [11,12], dipeptide composition [12,32], and other hybrid features [25]. Kumar et al [20] newly incorporated A C C E P T E D M A N U S C R I P T evolutionary information into sequence-based methods.…”

“…They addressed the data imbalance problem by a novel ensemble classifier (imDC; see [39]). Furthermore, they programmed their imDC classifier in an improved DNA-binding protein predictor based on the 188D sequence physicochemical features.…”

iEnhancer-PsedeKNC: Identification of enhancers and their subgroups based on Pseudo degenerate kmer nucleotide composition

“…In our future work, we will try other features considering long range sequence information, which is indicted to be useful for the enhancer classification in our current work. We will also try the imbalanced classifiers [52][53][54][55][56]on our dataset, which has been employed CD-HIT and random sampling strategy for the large negative data. Combined with some more sophisticated machine learning models and feature reduction methods [57], we anticipate better performance can be achieved.…”