A new version of FASTA 2.0u4 (for unix) is available. This
version fixes a bugs with LFASTA on very long sequences and corrects
behaviour with 12-character sequence identifiers.
Also note that FASTA 2.0u4 can read GCG format libraries (though
not query sequences), so sites running GCG can update to a version of
FASTA that provides accurate statistical estimates.