May 13, 1997

Version overview

Currently, the fasta2u65.shar.Z is the latest complete fasta package,
and it has a complete set of searching programs (fasta, ssearch,
fastx, etc).  However, the searching programs are more in a maintenance
mode, bug fixes only.

The fasta3 series, which has ONLY the searching programs, has the
latest versions of the algorithms and statistical methods.  fasta3
also runs the exact same functions threaded (fasta3, fasta3_t) and in
parallel using PVM.

Here is a list of the programs, and where they can be found:

program		fasta2		fasta3			replaced by

fasta		yes		fasta3, fasta3_t

ssearch		yes		ssearch3, ssearch3_t

tfasta		yes		tfasta3, tfasta3_t	(tfastx3 preferred)

fastx		yes		fastx3, fastx3_t

tfastx		yes		tfastx3, tfastx3_t

rdf2 (obsolete)	no		no			prdf2

rss  (obsolete)	no		no			prss

prdf2		yes		no

prss		yes		no

lfasta		yes		no

lalign		yes		no

plalign		yes		no

flalign		yes		no

align		yes		no

align0		yes		no

randseq		yes		no

crandseq	yes		no

aacomp		yes		no

bestscor	yes		no

fromgb		yes		no

grease		yes		no

tgrease		yes		no

garnier		yes		no

zs_exp		yes		no

Changes with version 1.5

	FASTA version 1.5 includes a number of substantial revisions
to improve the performance and sensistivity of the program. Two
changes are apparent.  It is now possible to tell the program to
optimize all of the init1 scores greater than a threshold.  The
threshold is set at the same value as the old FASTA cutoff score
(approximately 0.5 standard deviations above the mean for average
length sequences).  For highest sensitivity, you can use the -c option
to set the threshold to 1.  (This will slow the search down about
5-fold).  In addition, you can tell FASTA to sort the results by the
"init1" score, rather than the "initn" score, by using the "-1"
option.  FASTA -1 ... will report the results the way the older FASTP
program did.

	A new method has been provided for selecting libraries. In the
past, one could enter the name of a sequence file to be searched or a
single letter that would specify a library from the list included in
the $FASTLIBS file. Now, you can specify a set of library files with a
string of letters preceeded by a '%'.  Thus, if the FASTLIBS file has
the lines:

	Genbank 64 primates$1P/seqlib/gbpri.seq
	Genbank 64 rodents$1R/seqlib/gbrod.seq
	Genbank 64 other mammals$1M/seqlib/gbmam.seq
	Genbank 64 vertebrates $1B/seqlib/gbvrt.seq

Then the string: "%PRMB" would tell FASTA to search the four libraries
listed above.  The %PRMB string can be entered either on the command
line or when the program asks for a filename or library letter.

	FASTA1.5 also provides additional flexibility for specifying
the number of results and alignments to be displayed with the -Q
(quiet) option.  The "-b number" option allows you to specify the number of
sequence scores to show when the search is finished.  Thus

	FASTA -b 100 ...

would tell the program to display the top 100 sequence scores. In the
past, if you displayed 100 scores (in -Q mode), you would also have
store 100 alignments. The "-d" option allows you to limit the number
of alignments shown.  FASTA -b 100 -d 20 would show 100 scores and 20
alignments.

	The old "CUTOFF" parameter is no longer used.  The program
stores the best 2000 (IBM-PC, MAC) or 6000 (Unix, VMS) scores and then
throws out the lowest 25%, stores the next 500 (1500) better than the
threshold determined with the first scores were discarded, and repeats
the process as the library is scanned.  As a result, the best 1500 -
2000 (4500 - 6000) scores are saved.  The old cut-off parameter was
also used to set the joining threshold for the calculation of the
initn score from initial regions.  This joining threshold can now be
set with the -g option or the GAPCUT parameter.

	Finally, FASTA can provide a complete list of all of the
sequences and scores calculated to a file with the "-r" (results)
option.  FASTA -r results.out ... creates a file with a list of scores
for every sequence in the library.  The list is not sorted, and only
includes those scores calculated during the initial scan of the
library (the optimized score is not calculated unless the -o option is
used).

Changes with 1.6c31a

	(August, 1993) Released support for NCBI SEARCH and
	BLASTP/BLASTN formats.

	(November, 1993) Changes to nxgetaa.c to accomodate changes in
	embl library format. Changes to ncbl_lib.c to work on DNA
	sequences

Changes with 1.6c24

	(December 1992)  Added -e option for more selective scores.

	(April 1993) Added #define SUPERFAMNUM for genpept.fasta
	users.  By default, superfamily numbers are not returned from
	fasta format (libtype=1) files.

	(May 1993) Changed window shuffle routine in rdf2, rss, to
	preserve locality of shuffle.

Changes with version 1.6b

	FASTA version 1.6b uses a new method for calculating optimal
scores in a band (the optimization or last step in the FASTA
algorithm). In addition, it uses a linear-space method for calculating
the actual alignments.  The FASTA package also includes four new
programs:

	ssearch		a program to search a sequence database using
			the rigorous Smith-Waterman algorith (this
			program is about 100-fold slower than FASTA
			with ktup=2 (for proteins).

	rss		a version of rdf2 that uses a rigorous
			Smith-Waterman calculation to score
			similarities

	lalign		A rigorous local sequence alignment program
			that will display the N-best local alignments
			(N=10 by default).

	plalign		a version of lalign that plots the local alignments.


	The lalign/plalign program incorporate the "sim" algorithm
described by Huang and Miller (1991) Adv. Appl. Math. 12:337-357.
The ssearch and rss programs incorporate algorithms described by
Huang, Hardison, and Miller (1990) CABIOS 6:373-381.

	Two new command line options are available:

	-n	indicates that the query file is a nucleotide
		sequence.  This  option can be very useful when
		searching with consensus regulatory sequences.

	-x "off1 off2"  allows you to specify an offset for the
		beginning of a DNA or protein sequence.  For example,
		if you are comparing upstream regions for two genes, and
		the first sequence contains 500 nt of upstream
		sequence while the second contains 300 nt of upstream
		sequence, you might try:

		fasta -x "-500 -300" seq1.nt seq2.nt

		This option will not work properly with the translated
		library sequence with tfasta.

		(You should double check to be certain the negative
		numbering works properly.)

Changes with version 1.5

	FASTA version 1.5 includes a number of substantial revisions
to improve the performance and sensistivity of the program. Two
changes are apparent.  It is now possible to tell the program to
optimize all of the init1 scores greater than a threshold.  The
threshold is set at the same value as the old FASTA cutoff score
(approximately 0.5 standard deviations above the mean for average
length sequences).  For highest sensitivity, you can use the -c option
to set the threshold to 1.  (This will slow the search down about
5-fold).  In addition, you can tell FASTA to sort the results by the
"init1" score, rather than the "initn" score, by using the "-1"
option.  FASTA -1 ... will report the results the way the older FASTP
program did.

	A new method has been provided for selecting libraries. In the
past, one could enter the name of a sequence file to be searched or a
single letter that would specify a library from the list included in
the $FASTLIBS file. Now, you can specify a set of library files with a
string of letters preceeded by a '%'.  Thus, if the FASTLIBS file has
the lines:

	Genbank 64 primates$1P/seqlib/gbpri.seq
	Genbank 64 rodents$1R/seqlib/gbrod.seq
	Genbank 64 other mammals$1M/seqlib/gbmam.seq
	Genbank 64 vertebrates $1B/seqlib/gbvrt.seq

Then the string: "%PRMB" would tell FASTA to search the four libraries
listed above.  The %PRMB string can be entered either on the command
line or when the program asks for a filename or library letter.

	FASTA1.5 also provides additional flexibility for specifying
the number of results and alignments to be displayed with the -Q
(quiet) option.  The "-b number" option allows you to specify the number of
sequence scores to show when the search is finished.  Thus

	FASTA -b 100 ...

would tell the program to display the top 100 sequence scores. In the
past, if you displayed 100 scores (in -Q mode), you would also have
store 100 alignments. The "-d" option allows you to limit the number
of alignments shown.  FASTA -b 100 -d 20 would show 100 scores and 20
alignments.

	The old "CUTOFF" parameter is no longer used.  The program
stores the best 2000 (IBM-PC, MAC) or 6000 (Unix, VMS) scores and then
throws out the lowest 25%, stores the next 500 (1500) better than the
threshold determined with the first scores were discarded, and repeats
the process as the library is scanned.  As a result, the best 1500 -
2000 (4500 - 6000) scores are saved.  The old cut-off parameter was
also used to set the joining threshold for the calculation of the
initn score from initial regions.  This joining threshold can now be
set with the -g option or the GAPCUT parameter.

	Finally, FASTA can provide a complete list of all of the
sequences and scores calculated to a file with the "-r" (results)
option.  FASTA -r results.out ... creates a file with a list of scores
for every sequence in the library.  The list is not sorted, and only
includes those scores calculated during the initial scan of the
library (the optimized score is not calculated unless the -o option is
used).

Changes with 1.7

	(February 1994) Replaced rdf2 and rss with prdf, prss, which
	calculate informative score probabilities by fitting the
	distribution of shuffled scores to an extreme value
	distribution.  The curve fitting routines in rweibull.c were
	provided by Phil Green, Washington U., St. Louis.

	"-i" switch to reverse-complement query sequence if it is
	DNA.

	Fix bug in zzlgmata.c that caused problems with alignments.

	Fix histogram routine to work properly with ln() normalized
	scores.

Changes with 2.0  (March, 1995)

	WARNING - Optimization is now turned on by default.  The
	meaning of the "-o" option has been reversed.  "-o" now turns
	off optimization, reverting to the earlier method of sorting
	by "initn" scores.

	Change default protein matrix to BLOSUM50.  PAM250 is
	still available with -s 250.  Change program to accept gap
	penalties from the command line with "-f" (-12) and "-g" (-2).

	Provide MARKX=4, which allows one to display the conserved
	regions of the query sequence after a library search.

	Calculate explicit probability estimates for FASTA, TFASTA,
	and SSEARCH.  Estimates assume that the library contains a large
	number of unrelated sequences.  If this is not correct, the
	estimates are useless (and should be turned off with the -z
	flag).

	The width of the band used to calculate optimized scores is
	now variable.  For proteins and ktup=1, 32 residues are used,
	otherwise 16 residues are used.  For DNA, 16 residues are
	used. This value can be changed with the "-y" option.

	FASTA alignments now use the Smith-Waterman algorithm; there
	is no longer a limit on gap size for FASTA alignments.

	Fixed a rare bug in lalign/plalign for low gap penalties.

	Fixed lfasta to read one letter filenames in second position.

April 5, 1995

	Fixed bug in blast-format file reading treat sequences that do
	not end in "*" properly.

May, 1995

	More accurate display of the expected value histogram.  The
	quality of the fit is now quantitated with the
	Kolmogorov-Smirnov statistic.

	DNA match/mismatch penalties changed to +5/-4.

	An expectation theshold (-E) is provided for displaying
	scores.

July, 1995

	Corrected a very serious bug in ssearch E()-score calculation
	for large databases.

	Corrected a minor problem with histogram scaling.

	Removed Kolmogorov-Smirnov statistic if histogram not shown.

	Show correct scoring matrix if specified matrix is not
	found.

August, 1995

	Some corrections so that "-z" flag works properly and statistical
	calculations fall back properly when no distribution of lengths
	is available.

2.0x3	Change default DNA and TFASTA alignments to older band-limited
	Smith-Waterman rather than full Smith-Waterman.  Now DNA
	sequence searches are as fast as before (with Smith-Waterman
	alignments, they were often 50 times slower).  Full Smith-Waterman
	alignments are available with the "-A" option.

	Small changes in the way that memory is allocated for
	alignments in FASTA, TFASTA, LFASTA/PLFASTA, and SSEARCH.

	The DOS/BorlandC and UNIX versions have been merged.  All
	files necessary for compilation on Dos/WinNT are included.

	Added -O option to FASTA, TFASTA, LFASTA, PRSS, PRDF, LALIGN, ALIGN
	to specify output file.

2.0u3	merge of Mac FASTA with Win/DOS, Unix FASTA to a single set of files.

Sept, 1995

	add -Q option to prss, prdf.  Fix bug in -O option for those
	programs.

	Allow longer lengths for filenames.  Use QFILE_SIZE and LFILE_SIZE
	to define lengths for query and library file names (40, 80 for
	microcomputers, 256 for Unix).

November, 1995

	Fix bug in nxgetaa.c that prevented reading multiple
	blast-formatted files.

February, 1996

	see readme.v20u4 for more information

	added -m 10 option for parseable output

	added library_type 6 for GCG formatted files

	added -L option for long descriptions of library sequences

	"randseq" random shuffle program now available.

March, 1996

	modified nxgetaa for 12 character locus names.

	fixed a bug in lfasta that appears with very long sequences

April, 1996

	Make certain '-z' option really works (required for libraries with
	sequences < 10 aa).

	Removed duplicate sw_score: in ssearch with -m 10.

	Added -DPROGRESS to report progress of search with "....".

Mar, 1996

	Added "fastx", see readme.v20u5.  "-h" is not "-H".

Changes with 2.0u4 (February, 1996)

Added '-L' option, which provides a longer discription of the library
sequence.

Fixed a bug in the -m 10 parseable output.

Support is now provided for version 8.0 GCG libraries, both protein
and DNA. Use library type 6.

Changes with 2.0x4  (January, 1996)

The major change in with 2.0x4 is the ability to get a parseable
output from FASTA/TFASTA/SSEARCH.  This can be done using output
option -m 10.  With -m 10, the initial histogram and list of best
scores is unchanges, but the alignments are now in a parseable form:

>>>mgstm1.aa, 217 aa vs s library
; pg_name: FASTA
; pg_ver: version 2.0x4 Jan., 1996
; pg_matrix: BLOSUM50
; pg_gap-pen: -12 -2
; pg_ktup: 1
; pg_optcut: 30
; pg_cgap: 42
>>GTB1_MOUSE GLUTATHIONE S-TRANSFERASE GT8.7 (EC 2.5.1.18
; fa_initn: 1490
; fa_init1: 1490
; fa_opt: 1490
; fa_z-score: 1916.0
; fa_expect:      0
; sw_score: 1490
; sw_ident: 1.000
; sw_overlap: 217
>GT8.7  ..
; sq_len: 217
; sq_type: p
; al_start: 1
; al_stop: 217
; al_display_start: 1
PMILGYWNVRGLTHPIRMLLEYTDSSYDEKRYTMGDAPDFDRSQWLNEKF
KLGLDFPNLPYLIDGSHKITQSNAILRYLARKHHLDGETEEERIRADIVE
NQVMDTRMQLIMLCYNPDFEKQKPEFLKTIPEKMKLYSEFLGKRPWFAGD
KVTYVDFLAYDILDQYRMFEPKCLDAFPNLRDFLARFEGLKKISAYMKSS
RYIATPIFSKMAHWSNK
>GTB1_MOUSE ..
; sq_len: 217
; sq_type: p
; al_start: 1
; al_stop: 217
; al_display_start: 1
PMILGYWNVRGLTHPIRMLLEYTDSSYDEKRYTMGDAPDFDRSQWLNEKF
KLGLDFPNLPYLIDGSHKITQSNAILRYLARKHHLDGETEEERIRADIVE
NQVMDTRMQLIMLCYNPDFEKQKPEFLKTIPEKMKLYSEFLGKRPWFAGD
KVTYVDFLAYDILDQYRMFEPKCLDAFPNLRDFLARFEGLKKISAYMKSS
RYIATPIFSKMAHWSNK
>>GT28_SCHJA GLUTATHIONE S-TRANSFERASE 28 KD (EC 2.5.1.18
; fa_initn: 190
; fa_init1: 97
; fa_opt: 169
; fa_z-score: 217.9
; fa_expect: 1.1e-05
; sw_score: 169
; sw_ident: 0.277
; sw_overlap: 228
>GT8.7  ..
; sq_len: 217
; sq_type: p
; al_start: 4
; al_stop: 180
; al_display_start: 1
PMILGYWNVRGLTHPIRMLLEYTDSSYDEKRYTMGDAPDFDRSQWLNEKF
KLGLDFPNLPY--LID--GSHK-ITQSNAILRYLARKHHLDGETEEERIR
ADIVENQVMDTRMQLIMLCYNPDFEKQK--PEFLK-TIPEKMKLYSEFLG
KRP--WFAGDKVTYVDFLAYDILDQYRMFEPKCLDA-FPNLRDFLARFEG
LKKISAYMKSSRYIATPIFSKMAHWSNK
>GT28_SCHJA ..
; sq_len: 206
; sq_type: p
; al_start: 3
; al_stop: 180
; al_display_start: 1
-VKLIYFNGRGRAEPIRMILVAAGVEFEDERIEFQDWP----------KI
KPTIPGGRLPIVKITDKRGDVKTMSESLAIARFIARKHNMMGDTDDEYYI
IEKMIGQVEDVESEYHKTLIKPPEEKEKISKEILNGKVPILLQAICETLK
ESTGNLTVGDKVTLADVVLIASIDHITDLDKEFLTGKYPEIHKHRKHLLA
TSPKLAKYLSERHATAF
>>GT2_DROME GLUTATHIONE S-TRANSFERASE 2 (EC 2.5.1.18).
; fa_initn: 124
; fa_init1: 124
; fa_opt: 164
; fa_z-score: 210.1
; fa_expect: 2.9e-05
; sw_score: 164
; sw_ident: 0.248
; sw_overlap: 251
>GT8.7  ..
; sq_len: 217
; sq_type: p
; al_start: 4
; al_stop: 198
; al_display_start: 1
---------------------------PMILGYWNVRGLTHPIRMLLEYT
DSSYDEKRYTMGDAPDFDRSQWLNEKFKLGLDFPNLPYL-IDGSHKITQS
NAILRYLARKHHLDGETEEERIRADIVENQVMDTRMQLIMLCYNPDFEKQ
KPEFLKTIPEKMKLYSEFLGKR-----PWFAGDKVTYVDFLAYDILDQYR
-MFEPKCLDAFPNLRDFLARFEGLKKISAYMKSSRYIATPIFSKMAHWSN
K
>GT2_DROME ..
; sq_len: 247
; sq_type: p
; al_start: 52
; al_stop: 240
; al_display_start: 22
PPAEGAEGAVEGGEAAPPAEPAEPIKHSYTLFYFNVKALPSPC------A
TCSDGNQEYE--DVAHPRRVPALKPTMPMG----QMPVLEVDGK-RVHQS
ISMARFLAKTVGLCGATPWEDLQIDIVVDTINDFRLKIAVVSYEPEDEIK
EKKLVTLNAEVIPFYLEKLEQTVKDNDGHLALGKLTWADVYFAGITDYMN
YMVKRDLLEPYPAVRGVVDAVNALEPIKAWIEKRPVTEV


Note that the parseable output starts with ">>>" and that each
alignment record starts with ">>" while each aligned sequence record
starts with ">"

All parameters produced by the fasta package will be of the form:

	; xx_yyyyy

In this version, we have xx:

	pg - program parameters (name, version, matrix)
	fa - fasta scores, expect values, etc.
	sw - Smith-Waterman scores, expect values.
	sq - sequence length, type
	al - alignment start, stop, display_offset

Other FASTA distributors may choose to add additional fields.  If they
do, they should use a tag with more than two characters, e.g.:

	ebi_access:
or
	gcg_?????

The FASTA tags will be limited to two characters followed by a "_".

All of the output parameters correspond to values that are presented
in other FASTA output formats, with the exception of the "al_"
parameters.

al_start gives the location of the alignment start in the
	original sequence

al_stop gives the location of the end of the alignment in the
	original sequence

al_display_start
	gives the location of the first displayed amino acid residue
	in the original sequence.  The -m 10 alignments are the same
	as those produced in the other modes. In particular,
	FASTA/SSEARCH provide some context for the alignment; if the
	"-a" option is not used, FASTA/SSEARCH will try to provide
	about 30 residues on either side of the actual local
	alignment, if alignment is in the middle of one or the other
	sequence.  If the begining of the query sequence aligns with
	the 10'th residue of the library sequence, then the query
	sequence will be padded with ten leading "-" to produce the
	alignment.  The leading '-' are a formatting convenience only;
	they are not considered in the numbering system for
	al_display_start, al_start, or al_stop.

	Thus:

	>GT8.7  ..
	; sq_len: 217
	; sq_type: p
	; al_start: 3
	; al_stop: 180
	; al_display_start: 1
	---PMILGYWNVRGLTHPIRMLLEYTDSSYDEKRYTMGDAPDFDRSQWLN
	EKFKLGLDFPNLPYLIDGSHKITQSNAILRYLARKHH---LDGETEEERI
	RADIVENQVMDTRMQLIMLCYNPDFEKQKPEFLKTIPEKMKLYSEFLGKR
	PWFAGDKVTYVDFLAYDILDQYRMFEPKCLDA------FPNLRDFLARFE
	GLKKISAYMKSSRYIATPIFSKMAHWSNK
	>ARP2_TOBAC ..
	; sq_len: 223
	; sq_type: p
	; al_start: 6
	; al_stop: 181
	; al_display_start: 1
	MAEVKLLGFW-YSPFSHRVEWALKIKGVKYE---YIEEDRD--NKSSLLL
	QSNPV---YKKVPVLIHNGKPIVESMIILEYIDETFEGPSILPKDPYDRA
	LARFWAKFLDDKVAAVVNTFFRKGEEQEKGK--EEVYEMLKVLDNELKDK
	KFFAGDKFGFADIAANLVGFWLGVFEEGYGDVLVKSEKFPNFSKWRDEYI
	NCSQVNESLPPRDELLAFFRARFQAVVASRSAPK

 	Says that to align the two sequences, the first 'P' of GT8.7 must
	line up with the first 'V' (residue 4) in ARP2_TOBAC but that
	the actual best local alignment starts with the first 'I' in
	GT8.7 and the first 'L' in ARP2_TOBAC.

**Changes with release 20u5 - May 1996

This version of the FASTA package includes FASTX - a program that
compares a DNA sequence with a protein sequence library by translating
the DNA sequence in three frames and finding the best match, with
frame-shifts, between the translated DNA and protein sequence.
(unlike BLASTX, FASTX only does a three-frame translation.  To search
all six frames, do a second search with the "-i" option).

The code for aligning a three-frame protein sequence with a normal
protein sequence was provided by Zheng Zhang and W. Miller of the
Pennsylvania State University.

A third gap parameter, the frameshifts penalty, is provided with the
'-h' option.  The default gap penalties are -15 for the first residue
in a gap, -3 for each additional residue, and -30 for a frameshift.

The '-h' option used to prevent the histogram display, that option is
now invoked with '-H'.

Much of the FASTX code is new and has not been tested nearly as
extensively as the other fasta programs. Please inform me of bugs as
you find them.

================
**Changes with release 20u51 - June, 1996

Fixes to showalign for SHOWALL.

Fixes to routines that read fasta format files for long DNA sequences.

================
**Changes with release 20u52 - July, 1996

Fixes to lalign/plalign for setting gap penalties on DNA

Fixes to fastx to correct bug in alignment routine.

================
**Changes with release 20u53 - July, 1996

Another fix to fastx

Added flalign, a version of plalign that produces a GCG fig file for
local aligment graphics.

Increased the size of sequences that can be aligned by lalign to
120,000 residues with BIGMEM.

First release of Mac version with FASTX.

================

Bill Pearson
wrp@virginia.edu

** Changes with release 20u67 - May, 1999

Corrected serious problem with fastx alignments.

** Changes with release 20u66 - September, 1998

Made plalign, plfasta, psgrease, to be used with WWW pages.

plalign and plfasta now generate postscript graphics, rather
than tektronix graphics.

psgrease makes postscript Kyte-Doolittle plots.

Provide various *.htm and *.cgi files to implement WWW pages for
lalign, plalign, grease, chofas, garnier.

Updated grease (tgrease), chofas, and garnier for consistent
user interface.

** Changes with release 20u65 - May 1998

Various minor bug-tweeks to the fastx function, faatran.c, and other
programs associated with fastx.

** Changes with release 20u64 - May, 1998

Programs have been modified to accept query sequences from STDIN for
WWW interfaces.  prss, prdf, lalign, and plalign should accept input
from STDIN. This makes it relatively easy to set up a prss WWW site.

The translation routine used by FASTX has been modified to translate
ambiguous nucleotides as 'X'.

Problems with specifying gap penalties with DNA sequences have been
corrected in prss.

**Changes with release 20u6 - August 1996

Another new program - TFASTX - compares a protein sequence to a
translated, potenitally frameshifted, DNA library.  TFASTX is a
substantial improvement over TFASTA, although TFASTX is slower.

The LALIGN/PLALIGN family now includes FLALIGN, which will write out
alignment plots in GCG's Figure format.

Mac - version now uses System7 Standard File routines.

BLOSUM62 support finally included.

** September, 1996

Fixed another bug in fastx/tfastx.

** September, 1996

Fixed problem with query subsequence selection.

Fixed problem with selectbestz().

** September 23, 1996

Fixed problem with in -m=10 Smith-Waterman alignments pointed
out and corrected by Erik Wallen (erikw@biokemi.su.se).

**Changes with release 20u61 - November, 1996

Made corrections to fffasta.c, nxgetaa.c to support alternative
protein scoring matrices with fastx.

**Changes with release 20u62 - December, 1996

A fix to nxgetaa.c to allow -i with fastx (bug caused by 20u61
changes).

**Changes with release 20u63 - December, 1996

Corrected some problems with lfasta using -m 10 (parseable) output.

**Changes with release 20u64 - September, 1997

Modified faatran.c so that fastx searches with many "X"'s are
translated to 'X', not 'K'.

Added -m 5 (MARKX), which combines -m 0 and -m 4.

Corrected problem with lalign/plalign/flalign and
lfasta/plfasta/flfasta when reverse complemented DNA sequences were
compared to the same file.


Bill Pearson
wrp@virginia.edu

** January, 2007

Modify align.c (used for global alignments) to use the same scoring matrices, and other options, as lalign2.c.

Makefile has been modified so that make all only makes FASTA2 programs
that are not part of FASTA3.  The search programs are no longer made
by default.

** December, 2006

Modify fffasta.c to support the **pam2 global in upam.gbl, rather than
pam2[][] - **pam2 was introduced for lalign2.c/lsim3.c.

** August, October 2006, 21u08 (lalign2.c)

Make some efforts to remove global variables from lsim2.c, by
replacing it with lsim3.c.  Initial efforts in August, 2006,
introduced a bug, which was detected and fixed in October.

Provide option to show identical alignments.

** May, 2005,  21u07 (lalign.c)

Modify the code that checks for identical sequences to not assume
sequences are identical just because the filenames are.  They may
be different because of sub-setting.

Add -I option to show identical alignment.

Update lalign.1 documentation.

** April, 2004, 21u06 (lalign2.c)

Fix problem reading external scoring matrix files.  The file
was not read, and then sequences were not read properly.

Incorporate GAP_OPEN gap matrix options.

Changes to allow G:U RNA base matches.

** March, 2000, 21u02 (lalign.c)

Added '-N length' option to limit query, library sequences to
"length".  Corrected problems with sequence numbering when
subsequences were specified.

Modifications to nrand.c to keep more bits and return random numbers
from 0..n-1.  Use "nrandom.c" rather than nrand.c if random() is
available.

Fixes to shuffling routines in randlib.c.

** November, 2003

Add Blosum80 matrix to lalign.c, upam.gbl.