If I recall correctly, we required junctions that supported gene merges to be cross validated by another data source (cDNAs, RT-PCR, 454, etc.). I've attached a longer list which includes more junctions that were filtered out. We erred on the side of caution figuring that we had enough novel discoveries that were rigorous that we didn't want to dilute the study with less confident results. In any event, attached is the longer list. Feel free to use these, but I would ask that you only make public the extras in this list that you have high confidence in. Brenton R Graveley UConn Health Center