stop relying on specific contig names

[mschilli87]

Modifying seqlevels like here is not robust (in C. elegans for example UCSC's "chrM" is not "MT" in Ensembl, but "MtDNA" (or something like this)).

I'd suggest to do these kind of things by modifying seqlevelsStyle.

Maybe we can even get rid of this completely & filter out mitochondrial circRNAs without assuming a fixed contig name?

[al2na]

https://dl.dropboxusercontent.com/u/1373164/please_check_the_code_for_bugs..mp3