Design Of Optimization Algorithm For DNA Coding Sequence Detection

With the mass emergence of bio-genetic information, the detection of DNA coding sequence based on digital signal processing was proposed. Taking advantage of its speed and efficiency, detection method based on DNA coding sequence features are specifically designed and has been a significant development.Most traditional coding sequence detection algorithms rely solely on the existence of period-3 behavior of amplitude-frequency characteristics, and the zero-padding extension neither contributes to the power of amplitude-frequency characteristics, nor makes accurate correction of the location offset. Due to these, in this paper, we proposed an algorithm of period-3 coding sequence detection based on double-boundary periodic extension combining with C and G content, which enhanced the indicative performance of the detection curve and the robustness of the method.Also detection results were further discerned and optimized by the rules of the coding region of DNA sequences. Detection results at the nucleotide level rate evaluated by the ROC reached 86.87% on dataset HMR195, which confirms that the combination method further adjusts the location offset. Combined with C and G content and coding regulation, this algorithm outperforms existing AMDF, TDP and other time and frequency domain detection algorithms without increasing time complexity at the same time.Finally, based on the coding sequence features extracted in this article, support vector machine (SVM) classifier was introduced here. Testing on the data set HMR195 the DNA sequence at the nucleotide-level, the classifier generated a good result.