Hello all, Further to recent discussions about dealing with large FASTQ files, and the short term advantages of gzipped FASTQ as a filetype in Galaxy (or gzipped files in general, https://bitbucket.org/galaxy/galaxy-central/issue/666/ ), we also discussed the merits of unaligned BAM files. It occurred to me this could be defined in Galaxy as subclass of the existing BAM definition, with a simple sniffer based on running "samtools idxstats" to find out if there are any mapped reads or not. This would then make it easy for tools capable of aligning SAM/BAM to list unaligned-bam as a Galaxy input format (avoiding complications of realigning previously aligned reads). You could also have a matching aligned BAM subclass, then tools like SNP detectors should insist on an aligned BAM file for input. Does this sound reasonable? Peter