Changelog

#
#Version: 4.2.2
#OS:      Linux
#Author:  Sebastien Moretti
#E-mail:  moretti.sebastien [AT] gmail.com
#
4.2.2
	Use new and simpler blast submitter/parser

4.2.1
	Fix for revtrans option that cannot be strictly boolean for the new
	Tcoffee web server
	Follow NCBI changes for GI to acc.version, and https

4.2.0
	Switch to exonerate 2.2.0

4.0.7
	Harmonize download methods
	Deep cleaning of code and names
	Bug fixes: ...

3.4.2
	Correct download for GeneID: 726214 / LOC726214
	Enhance colored fasta CDS

3.4.1
	Enhance colored fasta CDS

3.4.0
	Add html/css output

3.2.6
	Remove '>' sign in annotation line if any e.g.: NM_105729

3.2.5
	The fasta file is too large, try with less than 60 sequences or
	split your file

3.2.4
	Add a simple way to get chromosomic sequences info about gene
	acc and positions via eSummary

3.2.3
	Add more equivalent hits to ProtoGene

3.2.2
	Use more Perl Best Practices
	NCBI links for NTUI has changed

3.2.1
	Do not count N letters in (1st) sequence to avoid that nt seq with many N
	(indetermined base) is considered as amino acid

3.2.0
	Test if Gigablaster is reachable/available then use it else use NCBI blast

3.0.4
	Correct fasta headers in advance mode for CDS & CDSP: elements were
	duplicated and a > was persistent.
	Add exception for warnings when empty lines or non-fasta file were
	used in ProtoGene advance mode

3.0.3
	Correct seq_name when starting with a blank
	e.g.: > truc
	Correct when first seq is revtrans and check for nt

3.0.2
	Print an error message when input files are already nucleotides

2.2.5
	Best management when only one gene match and alignment fails:
	$intronStep==1 and $POS->{0} doesn't exist.
	Best management of fake downloads:
	- > without sequences
	- sequences with Error: or <>
	- ...

2.2.4
	Change intronless behavior if best nucleotide return
	a GI number.

2.2.3
	Change gene locus coordinates to zero if they become
	negative with +/- 5000 expansion.

2.2.2
	Manage boj output when there are several intronless genes
	for one query.

2.2.1
	Add a $lim variable to be able to fix max number of query
	sequences as an input option.
	Manage several geneIDs ?!?

2.2.0
	New Release:
	Better memory management
	Code splits into more functions
	Clean code
	Write results when they are available, no more only at the exit

1.2.8
	Add an option to keep abnormal temporary files.
	Increase size threshold for result file of exonerate, i.e.
	no result or result.

1.2.6
	Manage fake BLASTP acc when they have space, point or pipe
	characters.
	Check and correct perl POD text format.
	Enhance fasta header for boj output like cds output.

1.2.4
	Add an option to hide BOJ output for the IGS web server.
	Output two CDS files: a standard one, and one with query prot
	interleaved.

1.2.2
	Manage multiple runs of PACMAN and the co-occurance query name:
	0.fas and 0.fas  => now use temp name to limit this.
	And temp names are mix of current date-time and a random nbr.

1.2.0
	Remove duplicated lines when a query, with multiple blastp hits
	returns a single nucleotide for all hits. So, keeps only one.
	Debug mode for BOJ, now available as a result file.
	Enhance output for log files.
	RefSeq -> 100% filters to avoid mis-species attribution
	then NR -> 95%.

1.0.4
	Add a webblast step againt nr if refseq webblast returns no
	hit: RefSeq, elsif nr, else NO HIT

1.0.3
	Add an option to add the original peptide query beneath the
	back-translated nt seq
	Add the b,o,j identification in the Exonerate parser module

1.0.2
	New header format in the cds output

1.0.1
	Add standard log files for the web server
	Manage multiple equal blast best hits
	Mix PACMAN and Exoset, specially in the Exonerate parser

0.99
	Process the alignment between every nucleotide sequence found and use
	the best one (transcript and genomic sequences)

0.98
	Add a 'old' argument for the cache option to remove only 15 days
	old files from the cache directory
	Hide webblast parameters when it runs but the first one
	Manage return lines from windows '^M'
	Remove all the non-gap and non-alphabetic characters into the query
	sequences

0.97
	Create a more robust 'cache directory' process
	Add triplet to sure mismatches, i.e. Met and Trp

0.96
	Add a second result file, as library, with storage of accession
	numbers resulting from blast query and prot-nt correspondances.

0.95 - 24 oct 2005
	Correct hash table variable define as semi-local, and not strictly as
	local, in Exonerate.pm because some value were kept between
	module executions due to misplaced brackets.
	Clean Exonerate.pm file and remove usage of strand, exonup
	list and genomic length because they are unnecessary here.

0.9 - 20 oct 2005
	Perldoc

19 oct 2005
	Addition of a test to avoid to download the same file if
	it has ever been downloaded in the past half day.
	Addition of stop codon triplet management
	Addition of a process to select the best alignment if there
	are several nt sequences of the same category for a query

18 oct 2005
	Get genomic and transcript acc, sort them to get only the
	best category of sequences:
	transcripts > chromosomes > contigs > bac,cosmid,... for RefSeqN
	sort by size for other acc

17 oct 2005
	Addition of a version option and variable