I have 3 BAM files of the same specie, each of ~7GB, from three experimental runs. I merged the three BAM files to produce a single 22 GB bam file, using samtools merge -r option. Then I sorted this merged bam file with samtools sort, and i got 11 GB merged bam. Is is possible to reduce the size of merged bam file by 50%??
When you sort by coordinate, you bring reads with similar sequences next to each other, allowing the compression algorithm to see more compressible content. It is worth checking, though, that the number of sequences is what you expect using `samtools flagstat` or simply `samtools view` and a `wc -l`.