Authors: I. Dubchak, S.R. Holbrook and I. Muchnik
Affilation: Lawrence Berkeley National Laboratory, United States
Pages: 38 - 41
Keywords: protein fold prediction, exon-intron discrimination, functional RNA genes, computer simulated neural networks
General global sequence descriptors have been developed which proved to be widely applicable to the prediction of properties of biological genes and gene products. These descriptors include composition, transition, and distribution of defined attributes in the amino acid or nucleotide sequence. We have tested this approach on three completely distinct biological problems: 1) prediction of protein three-dimensional folds, 2) discrimination between sequences of gene introns and exons, and 3) identification of putative RNA genes in genomic sequences.