A New, Dynamic-Representation-Based Gene Finding Method with an Analysis of False Positive Peaks

In this paper, we propose a new method for gene finding. The method uses a new dynamic representation scheme to map DNA sequences into a numerical form. The dynamic representation scheme assigns numerical pairs to the nucleotides based on their effectiveness in the period-3 spectrum. Nucleotides that have a stronger participation in the period-3 spectrum peaks are assigned numerical pairs that further enhance their participation. Another development that the proposed method introduces is the detection of the period-3 spectrum peaks to discriminate between protein coding and non-coding regions. In this paper, we also analyze the period-3 peaks that are predicted by the proposed method. We analyze the false positive peaks by scanning the stop codons in the possible reading frames. The work also analyzes the false positive peaks that are attached to true positive peaks. This analysis provides insights for future work that can be conducted to improve the prediction accuracy of spectrum-based techniques by studying the false positive peaks. In addition, it provides an insight about these false positive peaks that may have originated as transcribed sequences which, over time, acquired stop codons by mutation and lost their characteristic for transcription.