genemodel

A gene model is an annotated representation of a gene’s structure in a genome assembly. It specifies where a gene starts and ends, the locations of exons and introns, untranslated regions, and the coding sequence, as well as the strand on which the gene is encoded. A gene model may also include information about regulatory elements, transcript variants, and other features that help describe how the gene is transcribed and translated.

Gene models are produced by genome annotation efforts and are stored in major databases such as Ensembl, GENCODE, and RefSeq. They are generated from a combination of evidence sources, including transcript sequences (cDNA, ESTs, RNA-Seq), protein homology, and computational predictions. Ab initio gene predictors provide initial predictions, which are refined by evidence-based pipelines to produce more reliable annotations. Because many genes produce multiple transcripts through alternative splicing, gene models often include several isoforms with differing exon–intron structures.

In practice, gene models serve as coordinates for genome browsers, and as the basis for downstream analyses such as functional annotation, variant effect prediction, and comparative genomics. The accuracy of gene models depends on data quality and evidence integration; models are routinely revised as new transcriptomic and proteomic data become available. Challenges include distinguishing genuine genes from transcribed pseudogenes, precisely defining exon boundaries, and resolving complex loci with overlapping or nested genes. Advances in long-read sequencing and integrated annotation pipelines continue to improve the completeness and reliability of gene models across species.