Enhanced TAIR12 annotation (AraBase)

The official TAIR12 genome provides (1) a completed (T2T) assembly and (2) an updated and comprehensive annotation of the Arabidopsis thaliana genome . However, the original GFF3 file contains all annotation types (including protein-coding genes, noncoding RNAs, transposable elements, etc) in a single combined file. This structure can make it inconvenient to retrieve specific feature classes for downstream analyses. Moreover, the original protein-coding annotation did not contain explicit 5′ and 3′ UTR features. The organelle genomes (mitochondria and chloroplast) are missing in the TAIR12 files as well.

To improve accessibility and usability, the TAIR12 annotation was separated into multiple feature-specific GFF3 files. These include protein-coding genes, pseudogenes, noncoding RNAs, non-TE repeats, all transposable elements, Class I retrotransposons, Class II DNA transposons, and the major TE groups LTR, LINE, SINE, rolling-circle, and TIR elements. Separate genomic sequence files were also generated for all TEs, protein-coding genes, transcripts, CDSs, and predicted proteins.

To address the 5′ and 3′ UTR features missing issue, splice-aware UTR annotations were reconstructed for individual transcript isoforms using the exon and CDS structures in the GFF3 file. The predicted UTR intervals were then validated against the TAIR12 genome sequence and the officially distributed 5′ and 3′ UTR FASTA files. The resulting gene annotation now contains integrated five_prime_UTR and three_prime_UTR features, together with detailed validation and summary tables reporting sequence, length, chromosome, and coordinate agreement.

For the mitochondria and chloroplast chromosomes, we used the assemblies and annotations of TAIR10 version, which are completed assemblies, and merged to the TAIR12 genomes. The organelle genome annotations are in a separated file.

These processed TAIR12 resources have also been incorporated into AraBase, where users can browse and visualize genes, transcripts, UTRs, transposable elements, and other genomic features. AraBase additionally provides a convenient platform for comparing TAIR12 annotations with the widely used TAIR10/Araport11 annotation, helping users examine annotation updates, identify corresponding loci, and explore differences between genome releases.

All the scripts used in this process are provided, and the resulting files are available via Figshare. The AraBase are available via arabidopsis.ac.cn.

Please feel free to reach out if there are any questions.