Prediction of Protein Coding Regions by Support Vector Machine

Abstract

With the exponential growth of genomic sequences, there is an increasing demand to accurately identify protein coding regions from genomic sequences. Despite many progresses being made in the identification of protein coding regions by computational methods during recent years, the performances and efficiencies of the prediction methods still need to be improved. A novel method to predict the position of coding regions is proposed. First, a support vector machine is used as a classifier to recognize the first nucleotide of a codon in a coding region. Then, according to the difference of the time frequency characteristics of the output values of the classifier analyzed by Short Time Fourier Transform, the position of coding regions can be accurately determinate. The algorithm is not only can predict coding regions, but also can identify the first nucleotide of the codon in coding regions. This is very significant for accurate translation into a protein sequence. The simulation results show the proposed method is more effective for coding regions prediction than the existing coding region discovery tools.

A Review of Computational Intelligence Methods for Eukaryotic Promoter Prediction

“…The candidate features are determined using Markov-chain Monte Carlo (MCMC) sampler. ANN [42] Optimal Weight Vector ---TDNN [43] Weight Matrix ---FFNN [44] Data Window Slide -0.66 0.69 BPNN [7] Sliding Window ---ANN [2] Orthogonal Encoding -0.97 -Naive Bayes [2] Orthogonal Encoding -0.63 -SVM [23] -7-cross validation 0.80 0.82 SVM [48] Binary Feature Mapping Rule -0.99 0.99 SVM [52] Sparse Encoding 5-cross validation 0.96 0.95 FFBPNN [45] RS (Random sub-sampling) 5-cross validation 0.77 0.76 RVM [61] MCMC (Markov Chain Monte Carlo) sampler -0.60 0.56 SVM [54] Base-stacking energy of dinucleotides 5-cross validation 0.67 0.77…”

“…[5,50] In bioinformatics, applications of SVM includes splice site prediction, promoter prediction, host-pathogen separation. [51][52][53] The method explained in [22] uses statistical data for classification of promoter instead of features like TATA boxes, CAAT boxes and CpG islands. Sensitvity and specificity values attained using tetramer frequencies of bases is above 80%, still there is an opportunity for enhancement in the accuracy obtained.…”

“…[52] In this method, a sliding window of 91 nucleotides is given as input to SVM classifier after being encoded. Another attempt to utilize base stacking energy property of DNA has been made in.…”

Enhancing the prediction of protein coding regions in biological sequence via a deep learning framework with hybrid encoding★,★★

“…Meanwhile, the sliding window size is relevant to prediction performance, the use of a smaller size can increase the accuracy on the border of exon but leads to of higher rate of false positive prediction Hatzigeorgiou et al (1996). They Hatzigeorgiou et al (1996); Shuo and Yi-sheng (2009) use 91 whereas we adopt 90 in practice for convenience of counting the number of codons. The slight difference has almost no effect on the prediction performance.…”

Spectral Analysis of Exons in DNA Signals

“…It is a common phenomenon that nucleotide sequences in DNA perform a period three property [3,11] due to codon composition and structure in the strand. This fundamental characteristic can be exploited to predict the codon regions that help in determination of RNA sequences in DNA.…”