Tophat2 output bam size discrepancies

[Sentinel156]

Hi all,

I'm seeing tophat2 output bam files of different sizes (~800mb vs ~300mb) for two fastq files from the same strain (different biological replicates) with similar total read numbers for each (~14 million from 100bp single end illumina hiseq 2000 run). both fastqs are being mapped against the same reference genome but each sample was sequenced at different times and on different lanes. I don't believe this is a problem but would like to understand the size discrepancy.

Cheers,

[Brian Bushnell]

Things like mapping rate and error rate are much more useful than bam file size, which depends on such factors as compression level, read names, read quality score smoothness, and specific aligner parameters like which tags to generate and whether to output unmapped reads. Do you know the alignment rates and command line parameters?