Abstract
This paper provides a deterministic algorithm for finding a longest open reading frame (ORF) among all alternative splicings of a given DNA sequence. Finding protein encoding regions is a fundamental problem in genomic DNA sequence analysis and long ORFs generally provide good predictions of such regions. Although the number of splice variants is exponential in the number of optionally spliced regions, we are able to in many cases obtain quadratic or even linear performance. This efficiency is achieved by limiting the size of the search space for potential ORFs: by properly pruning the search space we can reduce the number of frames considered at any one time while guaranteeing that a longest open reading frame must be among the considered frames.