This directory contains real start codons and faked start codons,
where the faked ATG sites were selected from a window of +/- 100bp
around the actual start codon
This data is from
ftp://www-hgc.lbl.gov/pub/genesets/Drosophila/multi_exon_GB.sets.

This data set was created to build different splice site models.

	ATGreal_100_100.fa.gz	start codon sites
	ATGfaked_100_100.fa.gz	faked "ATG" sites


Both start codon data sets 100bp of the upstream region and 
100bp of the downstream region.

===================
Martin Reese, 26apr99 mgreese@lbl.gov