TopHat 1.3.0 release 6/2/2011
This release includes some very substantial performance improvements, bug fixes, and enhancements.
Geo Pertea has restructured the internal workflow of TopHat so that nearly all temporary data is compressed during a run. This greatly reduces the on-disk footprint of TopHat runs, and also reduces the amount of disk I/O. These enhancements should improve running time on networked filesystems.
Thanks to contributions from the Picard team (notably Alec Wysoker at the Broad Institute), TopHat has some needed SAM compliance fixes and some additional command line options that make it easier to incorporate TopHat into sequencing core workflows and automated processing pipelines.
TopHat can optionally use Bowtie's -n flag as an alternative to the read mapping protocol. The default is still to use -v, which can result in better spliced alignment accuracy when reads contain relatively few sequencing errors (particularly at the 3' end)
TopHat now supports both GTF2 and GFF3
TopHat can take gzip or bzip2 compressed FASTQ or FASTA files as input, decompressing them on the fly (the decompression program is determined based on the file name suffix)
The initial read inspection stage, used to determine the format, read length and other parameters needed for the run, is now significantly faster.
A bug in the installer caused the tophat script to be duplicated (apparently without causing further problems). Thanks to John Marshall for the fix to this and several other issues.
Numerous other minor bug fixes