[BioC] Varied performance of readGappedAlignmentPairs function

[Pan Du]

Hi Herve

I am using readGappedAlignmentPairs to retrieve the paired read information from BAM files. I found the performance of the function changed a lot from sample to sample, and it seems not related with the size of BAM files. 
 Also the increase of running time is not linear with the increase of transcript number, from which we want to retrieve information. Michael suggests it may be related with how readGappedAlignmentPairs function responds to weird alignments. I want to know whether you also observed similar problems and whether there is a solution for it. 
Thanks!

Pan