Until recently, the most common cost-effective way to discover new splice
variants was by aligning expressed sequences (ESTs and cDNAs) to the genome.
We developed a highly specific method for computational discovery of
alternative splicing, which is not based on EST analysis. This method
allowed us to discover hundreds of novel alternative splicing events not
documented so far. In the talk I will present our findings, including
additional innovative computational methods that are currently under
development.