Check slow splits

[joshfactorial]

No description provided.

[joshfactorial]

I'm tweaking when we use bgzip, which is quite slow. But I stumbled upon another issue, which is that the bam alignments are all showing up on the first 500,000 bases in the bam (when viewing in the IGV genome browser), so there must be an index off somwhere. I checkd chr1 of rice and the vcf index goes up to 44.3 million, which is roughly the size of chr1, so I think we are good there. Something still in the coordinates of the bam output.

[joshfactorial]

I made a small change to how the files were written, to try to speed up intermediate file creation. While testing that, I found an error in the way bams were indexing reads, causing only 500,000 bases worth of reads to get written. Now it spans the chromosome, but the coverage looks off.