gencode

GENCODE is a community-driven project that provides comprehensive annotation of gene features for the human and mouse reference genomes. The primary goal is to identify and curate all protein-coding genes, non-coding RNA genes, and pseudogenes, and to define the full set of transcripts and splice isoforms for each locus. Annotation includes gene structures, transcript coordinates, exon-intron boundaries, and metadata such as biotype and evidence.

The project combines automated computational annotation with manual curation. The manual component is carried out by the HAVANA group (Human and Vertebrate Analysis and Annotation) in collaboration with the broader GENCODE consortium. Evidence sources include RNA-seq, proteomics, cross-species conservation, and ongoing experimental validation. Each transcript receives a biotype label (for example protein_coding, lncRNA, antisense, pseudogene) and a transcript support level reflecting the confidence of the annotation.

Data products from GENCODE include GTF/GFF3 annotation files, transcript and gene identifiers (such as ENSG/ENST IDs used by Ensembl), and corresponding reference sequences. GENCODE annotations support multiple genome assemblies (e.g., GRCh38/hg38 in humans and GRCm38/mm10 in mouse) and are continuously updated through versioned releases. Data are disseminated via major genome browsers and databases, and are widely used for read alignment, gene and isoform expression analyses, variant effect prediction, and integrative genomic studies.

GENCODE is closely linked to the ENCODE project and forms a foundational resource for gene-level annotation in genomic research.