Changes and additions from ASTRAL 1.71 to 1.73:

* Bug fix on September 11-19, 2008:

The RAF file originally released with ASTRAL 1.73 contained incorrect
datestamps: the last REVDAT given in the PDB files was used instead of
the datestamp of the source file (as documented in the RAF
specification).  In addition, many RAF entries (mostly from 1.73, but
also a few manually edited entries from earlier releases) had the "one
to one" flag set incorrectly.  No other data were affected by this
bug, as these fields are used only for archival purposes and are not
used in generating other ASTRAL datasets.


* Bug fix on March 11, 2008:

ASTEROIDS versions released between Feb 18 and March 11 contained
errors (BLAST and Pfam matches to existing domains were not
generated).  The bug has been fixed, and corrected files were released
on March 17 (with weekly updates).


* Bug fix on Feb 7, 2008:

Some sequences containing more than 20% unknown residues
('x' characters) were accidently included in the sequence sets
corresponding to whole PDB chains and SCOP domains for classes 8 and
above.  No sequences in the standard sequence sets were affected by
the bug.  The sequences with more than 20% unknown residues have been
moved to the -reject files, as documented in the 2000 NAR paper.

The files that changed are:

scopseq-1.73/astral-chain-atom-all-1.73.fa
scopseq-1.73/astral-chain-atom-reject-1.73.fa
scopseq-1.73/astral-chain-seqres-all-1.73.fa
scopseq-1.73/astral-chain-seqres-reject-1.73.fa
scopseq-1.73/astral-chain-seqres-sel-gs-bib-100-1.73.fa
scopseq-1.73/astral-chain-seqres-sel-gs-bib-100-1.73.id
scopseq-1.73/astral-chain-seqres-sel-gs-bib-90-1.73.fa
scopseq-1.73/astral-chain-seqres-sel-gs-bib-90-1.73.id
scopseq-1.73/astral-chain-seqres-sel-gs-bib-95-1.73.fa
scopseq-1.73/astral-chain-seqres-sel-gs-bib-95-1.73.id
scopseq-1.73/astral-chain-seqres-sel-gs-bib-verbose-100-1.73.txt
scopseq-1.73/astral-chain-seqres-sel-gs-bib-verbose-90-1.73.txt
scopseq-1.73/astral-chain-seqres-sel-gs-bib-verbose-95-1.73.txt
scopseq-1.73/astral-ntcscopdom-seqres-gd-all-1.73.fa
scopseq-1.73/astral-ntcscopdom-seqres-gd-reject-1.73.fa

In the files above, sequences with more than 20% unknown residues were
moved from the -all files to the -reject files, and eliminated from
the selected subsets.  There was no case where the bug resulted in the
addition of a sequence to any set.


* Remediated PDB files

All sequences are now based on the remediated PDB files.  In many
cases, the chain id changed from a space to the letter A, which
changed the name of the domain in SCOP and ASTRAL (e.g., d101m__
changed to d101ma_).


* New RAF building procedure

The remediated PDB data set includes XML files that reliably map the
correspondence between SEQRES and ATOM residues, and also accurate
data on the original identity of post-translationally modified
residues (these are identified using the XML files and the chemical
dictionary).  This has resulted in a much lower number of chains that
require manual editing to correct errors in the automated translation:
from 915 chains in 1.71 to 23 chains in 1.73.  RAF is now built
directly from the XML files using a new scipt, xml2raf, which we plan
to release under an open source license.  The program previously used
to build RAF files from PDB format files, MakeRAF, has now been
deprecated.


* SPACI and AEROSPACI scores

We re-calculated AEROSPACI and SPACI scores for all PDB files in
ASTRAL 1.73, including entries that were present in older versions.
This was done in case PDB remediation affected the SPACI scores.


* ASTEROIDS

Pfam release 22 was used to create ASTEROIDS in ASTRAL 1.73;
these were first released on Feb 19, 2008.

We created ASTEROIDS using all protein chains from PDB entries that
are not classified in SCOP 1.73.  This includes entries released
before the SCOP freeze date (September 26, 2007), that were not
classified in SCOP, as well as entries released after the freeze date.
Nucleic acid chains, and additional chains from PDB entries that are
classified in SCOP, are not included in ASTEROIDS.