Automated annotation updates using PASA

The Program to Assemble Spliced Alignments (PASA) was developed and employed towards the incorporation of EST and FL-cDNA alignments into the TIGR Arabidopsis genome annotation. PASA was used to assemble 182,540 high quality EST and full-length cDNA (FL-cDNA) alignments (generated by blat, sim4, or geneseqer) into 25,165 alignment assemblies. These alignment assemblies were used as templates for updating gene structures and adding new genes to the Arabidopsis genome annotation, including the annotation of many alternative splicing variations, described in Haas et al, 2003. The following resources were developed as part of this effort:

• Alternative splicing variations


• Non-consensus splice sites


• PASA tool written in Perl. Single-exon EST alignments are allowed ambiguous orientations.


• Core of PASA, written in C++. This is a stripped down version of PASA which treats all alignments equivalently.



• The latest version of the complete PASA pipeline. Given a set of transcript sequences, genomic sequences, and gene structure annotations, the alignments are performed, assembled, and compared to the genome annotations. Test data sets are included as well as the CGI-scripts which enable web-based navigation of the results.


• 182,540 high quality EST and FL-cDNA spliced alignments to the Arabidopsis genome. The corresponding genomic sequences and transcript sequences are available.


• Resulting 25,165 alignment assemblies created using PASA