GARLIC V1.5

Changes:

• "unitialized value" warnings caused by UCSC model lookups fixed.
• TRF data should be in UCSC BED format which uses 0-based/half-open coordinates.
  Fixed a bug in the code where 1-based was assumed, causing a zero-valued start
  coordinate to go negative.
• createModel.pl: Added support for relative paths in input parameters.

GARLIC V1.4

• createModel.pl Fix: FASTA files with descriptions were not being parsed correctly.

GARLIC V1.3

createFakeSequence Changes:

• New option "--useBED" for createFakeSequence.pl. If this option is used,
  the insert file format is switched to a basic BED format.
• The inserts file now contains all inserts as discrete ranges rather than
  lumping together all recursive elements as one entry.
• Backup pre-existing *.inserts and *.fasta files rather than appending
  to them.
• Indicate in the header of the *.inserts file that the coordinates are
  zero-based half-open.
• The switch in createModel.pl from using RepeatMasker's *.out to *.align
  file for TE data removes the need to rename some element IDs in
  createFakeSequence.pl. ie. s/_\dend//; and s/-int//;

createModel Changes:

• Use RepeatMasker *.align file instead of *.out file to source repeats.
  This has the advantage of 1-1 mapping of annotatation names with consensus
  sequence names.

GARLIC V1.2

Workaround for bug in perl involving the use of very large strings. Only a problem when a user wants to generate a genome sized simulated sequence.

GARLIC V1.1

Changes include:

• The number of GC bins in the model is now an user changeable option in createModel.pl. The "-n #" parameter will create bins with GC ranges that equally divide the input genome.
• createModel.pl and createFakeSequence.pl speed improvements.
• Fix to code which classified the background GC for loci where repeats are inserted into other repeats.