Signal Processing in Sequence Analysis: Advances in Eukaryotic Gene Prediction

Genomic sequence processing has been an active area of research for the past two decades and has increasingly attracted the attention of digital signal processing researchers in recent years. A challenging open problem in deoxyribonucleic acid (DNA) sequence analysis is maximizing the prediction accuracy of eukaryotic gene locations and thereby protein coding regions. In this paper, DNA symbolic-to-numeric representations are presented and compared with existing techniques in terms of relative accuracy for the gene and exon prediction problem. Novel signal processing-based gene and exon prediction methods are then evaluated together with existing approaches at a nucleotide level using the Burset/Guigo1996, HMR195, and GENSCAN standard genomic datasets. A new technique for the recognition of acceptor splice sites is then proposed, which combines signal processing-based gene and exon prediction methods with an existing data-driven statistical method. By comparison with the acceptor splice site detection method used in the gene-finding program GENSCAN, the proposed DSP-statistical hybrid technique reveals a consistent reduction in false positives at different levels of sensitivity, averaging a 43% reduction when evaluated on the GENSCAN test set.

I. Introduction

Deoxyribonucleic acid (DNA), the material of heredity in most living organisms, consists of genic and intergenic regions, as shown in Fig. 1. In eukaryotes, genes are further divided into relatively small protein coding segments known as exons, interrupted by noncoding spacers known as introns. In eukaryotes such as human, the intergenic and intronic regions often make up more than 95% of their genomes. Codons (i.e., triplets of possible four types of DNA nucleotides , , , and ) in exons encode 20 amino acids and 3 terminator signals, known as stop codons (i.e., TAA, TAG, and TGA). Initial exons of the genes begin with a start codon “ATG.” Looking from the end of DNA (upstream) to its end (downstream), the exon-to-intron border is known as the donor splice site and consists of a consensus dinucleotide “GT” as the first two nucleotides of the intron, whereas the intron-to-exon border is known as the acceptor splice site, which consists of a consensus dinucleotide “AG” as the last two nucleotides of the intron. The accurate identification of genomic protein coding regions, along with the recognition of other signals and/or regions (shown in Fig. 1) would result in an ideal gene finding and annotation system.