2016-12-20 Release 1.2.4 (AndreaGiacomo, r317)

	- made code gcc-6 compliant

	(no new command-line flags)

2016-07-20 Release 1.2.3 (AndreaGiacomo, r313)

	- fixed consensus line in ClustalW format, 'H' in NHQK not recognised as strong (bug in Squid)

	(no new command-line flags)
	
2016-07-01 Release 1.2.2 (AndreaGiacomo, r310)

	- fixed a memory leak in hhalign
	- parallelised hhalign
	- increased MAC-RAM default from 2GB to 8GB

	(no new command-line flags)

2014-02-28 Release 1.2.1 (AndreaGiacomo, r292)

	- corrected Clustal format consensus line
	- label justification (multi-bit)
	
2013-06-12 Release 1.2.0 (AndreaGiacomo, r284)

	--is-profile              disable check if profile, force profile (default no)
	--use-kimura              use Kimura distance correction for aligned sequences (default no)
	--percent-id              convert distances into percent identities (default no)

	fixed bug in --output-order=tree-order for pair-wise alignment

	Kimura distance correction no longer default for protein
	
	fixed bug in Viterbi algorithm (predominantly affected long nucleotide but also protein, to lesser extent)

		Viterbi is less (~10x) memory hungry than MAC but still scales quadratically with length:

		2GB of RAM (ClustalO default) can align sequences of 6.5k residues with MAC and 18k with Viterbi

		8GB of RAM (EBI limit) can align sequences of 12.5k residues with MAC and 37k with Viterbi

		A Viterbi alignment of sequences of 100k residues will require 59GB of RAM
	
	
2013-05-16 Release 1.1.1 (r281)

	  --cluster-size=<n>        soft maximum of sequences in sub-clusters
	  --clustering-out=<file>   Clustering output file
	  --residuenumber, --resno  in Clustal format print residue numbers (default no)
	  --wrap=<n>                number of residues before line-wrap in output
	  --output-order={input-order,tree-order} MSA output order like in input/guide-tree

	turned off Kimura correction for DNA/RNA

	enable distance matrix output after initial mBed but subsequent full distance calculation

	enable termination after distance/guide-tree calculation (--max-hmm-iterations=-1)

	longer sequence labels 

2012-04-25 Release 1.1.0 (FilumVitae, r275)

   DNA and RNA support now added. Sequence type can be specified manually
   using --seqtype={Protein|DNA|RNA}

2012-03-27 Release 1.0.4 (r273)

   zipped input now supported

2011-09-07 Release 1.0.3 (r263) 

Bugs fixed:

   input failed if first line in fasta file was empty

   input failed if Clustal formatted file had trailing residue numbers

   '*' character was causing problems, did not get filtered out by squid 

   --outfmt=fasta was not recognized

   '~' gap characters were not recognized in MSF format

   amended README re sequence/profile alignment

   disallowed empty sequences

   doxygen documentation fixes and fix of API example

   introduced MAC-RAM flag to set amount of RAM given to MAC (in MB)
	
2011-06-23 Release 1.0.2

2011-06-17 Release 1.0.0 (DeshilHollesEamus, r250)

2011-04-14 Release 0.2.0 (IbiMeVidebunt, r238)

2011-03-23 Release 0.1.1 (OppidumLeonis, r224)
	
2011-02-22 Release 0.1.0 (OppidumLeonis, r200)
	
2010-08-10 Release 0.0.3 (Thamesis, r120)
	
2010-06-17 Release 0.0.1 (Dubliniensis, r81)
		
		First "release" as program has been able for a while to perform
		all basic tasks without problems.

		Non-standard features already built-in include:
		HMM-iteration (using HMMER for building an HMM) and guide-tree
		iteration. On top of that HMM input works fine and
		background-frequencies are added to the HHalign process.
		
		Known issues: RNA/DNA alignment is considered buggy. Aligned
		sequences have to be dealigned for HHalign to work properly.
		
		The HMMER version message can be ignored if no HMM-iteration
		was used.