Using a reference annotation combines what is known with what is novel
in your experiment. The "Common uses" section of this web page
this concept: http://cufflinks.cbcb.umd.edu/tutorial.html
The UCSC Known Genes track incorporates all of RefSeq Genes with other
sources, giving it an advantage. But it is also created at a
time (updated recently, check UCSC for exact dates for any track's
update). Tracks from Genbank are updated daily, includes RefSeq Genes,
giving it an different advantage as time passes. One or both tracks
or may not be available, depending on the target genome.
Another choice are the iGenomes builds at the Cufflinks web site for
UCSC. The iGenomes 'genes.gtf. files have specific attributes
that activate the full functionality of Cufflinks and related tools.
These are: gene_id, transcript_id, gene_name, p_id, and tss_id. For an
example of how these are used, see Cuffdiff usage, as explained here
. The iGenomes data
are a popular choice and were recently updated.
Note: mm9 'genes.gtf' is in Shared Data Libraries on Galaxy main.
Others are welcome to add to the post to share other good sources for
GTF datasets that include the attributes this tool package can