A grant effort is underway to sequence the entire human genome. Identification of functional sequence elements by computational tools has become increasingly important. Our long time goal is to use mathematical and statistical methods to identify most important functional elements-- protein coding genes and their regulatory elements from a large genomic sequences. Recently, we have developed a new human exon-finder called MZEF (PNAS 94:565-568, 1997), based on a modern multivariate statistical pattern recognition technique. In the proposed human chromosome 18 sequencing project, we plan to improve and to extend our MZEF in a real large-scale sequence production setting and to integrate new methods (especially first and last exon prediction methods) to be developed under the other funded research projects in Dr. Zhang's Laboratory. We shall interact with the informatics-core and other components closedly in order to test for accuracy, efficiency and scalability. Developing computational methods to assist rapid annotation of the gnome will also be vital for achieving the goals of Human Genome Project, which in turn will have a profound impact on human health and medicine.