Complete Gene Structure Prediction

Washington University, Saint Louis MO

Investigators

Abstract

Many genomes have already been sequenced and more will follow. However, for many research questions, the most important information in a genome is its translation products, not simply the nucleic acid sequence. Despite significant effort in experimental biology, high throughput sequencing, and bioinformatics, it is still difficult to answer the question, "What are the structures of all the genes and the amino acid sequences of the translation products?". The work here will use orthologous genomic sequences to assist in annotating the genomic sequence of other organisms. Specifically, the already assembled sequence of yeast, fruit fly, flat worm, and the first plant sequence, Arabidopsis, and orthologous shotgun sequence or EST sequence to help in identifying new gene structures. The system uses a probability model based on Hidden Markov Models and an optimization algorithm. This is a generalization of a project for comparing assembled orthologous sequences.