Optimal DNA-protein alignments with application to large-scale genome analysis

DNA-protein alignment algorithms can be used to discover coding sequences in a genomic sequence, if the corresponding protein derivatives are known. They can also be used to identify potential coding sequences of a newly sequenced genome by using proteins from related species. Previously known algorithms for computing DNA-protein alignments have one or more of the following drawbacks: not taking into account all aspects in problem formulation, providing optimal solutions that are run-time/memory expensive, and sacrificing optimality to achieve practical implementation. In this thesis, we present a comprehensive formulation of the DNA-protein alignment problem including indels, substitutions, frameshift errors, and intronic insertions between and within codons. We then provide an algorithm to compute an optimal alignment in O(mn) time using only four dynamic programming tables of size (m+1)x(n+1), where m and n are the lengths of the DNA and protein sequences, respectively. We developed a Protein and DNA Alignment program (PanDA) that implements the proposed solution. Experimental results indicate that our algorithm provides alignments that accurately reproduce GenBank annotation in nearly all cases when tested on gene and protein sequences from the same organism. We also present experimental evidence that our algorithm produces high-quality alignments and exon-intron predictions when aligning DNA sequences with proteins corresponding to orthologous genes from other species. We also present a parallel software that can be used to annotate, validate, and improve the quality of an assembly of a genome in a large scale. Spliced alignments between DNA sequences of the assembly and protein sequences from other organisms are done to achieve the same. Experimental results indicate that our software can produce putative annotations, while detecting candidate contigs to improve quality of an assembly