Title | EAnnot: a genome annotation tool using experimental evidence. |
Publication Type | Journal Article |
Year of Publication | 2004 |
Authors | Ding, L, Sabo, A, Berkowicz, N, Meyer, RR, Shotland, Y, Johnson, MR, Pepin, KH, Wilson, RK, Spieth, J |
Journal | Genome Res |
Volume | 14 |
Issue | 12 |
Pagination | 2503-9 |
Date Published | 2004 Dec |
ISSN | 1088-9051 |
Keywords | Algorithms, Base Sequence, Chromosomes, Human, Pair 6, Computational Biology, Genome, Genomics, Humans, Models, Genetic, Sensitivity and Specificity, Sequence Alignment |
Abstract | The sequence of any genome becomes most useful for biological experimentation when a complete and accurate gene set is available. Gene prediction programs offer an efficient way to generate an automated gene set. Manual annotation, when performed by experienced annotators, is more accurate and complete than automated annotation. However, it is a laborious and expensive process, and by its nature, introduces a degree of variability not found with automated annotation. EAnnot (Electronic Annotation) is a program originally developed for manually annotating the human genome. It combines the latest bioinformatics tools to extract and analyze a wide range of publicly available data in order to achieve fast and reliable automatic gene prediction and annotation. EAnnot builds gene models based on mRNA, EST, and protein alignments to genomic sequence, attaches supporting evidence to the corresponding genes, identifies pseudogenes, and locates poly(A) sites and signals. Here, we compare manual annotation of human chromosome 6 with annotation performed by EAnnot in order to assess the latter's accuracy. EAnnot can readily be applied to manual annotation of other eukaryotic genomes and can be used to rapidly obtain an automated gene set.
|
DOI | 10.1101/gr.3152604 |
Alternate Journal | Genome Res |
PubMed ID | 15574829 |
PubMed Central ID | PMC534675 |
Grant List | HG002042 / HG / NHGRI NIH HHS / United States |