To determine the impact of the intron-exon architecture and splice site strength on splice site selection, we created a database of alternative splice sites (ALTssDB) using the Human Exon Splicing Event Database HEXEvent, the Intron DB, and GeneBase. MaxEntScan, a computational tool was used to assign splice site scores. To minimize variability, we focused on competing alternative 5′ or 3′ splice site pairs of internal exons with only one alternative splice pattern. Thus, two alternative 5′ splice sites compete for a common 3′ss, or two alternative 3′ splice sites compete for a common 5′ss. As a result, ALTssDB reports the location of the major splice site and its competing alternative 3′ or 5′ splice site, corresponding exon sizes, usage levels, splice site scores and flanking intron lengths.

ALTssDB was created using EST data from the Human Exon splicing Events (HEXEvent) database. HEXEvent contains information regarding the location of competing splice sites, the resulting exon sizes, alternative splice site usage levels and the gene associated with each mRNA. The HEXEvent data was filtered to obtain a dataset comprising of only pairs of competing 5' and 3' splice sites separately. This database was subsequently modified to include splice site junction information and MaxENT scores using MaxEntScan. Using an R script and IntronDB dataset, (a database detailing eukaryotic intron features) flanking intron lengths were added to the database. Alternative splicing events were further filtered to include only events that have 10 or more EST counts.