Multiple EC_number tags for CDSs, lower case initials in /product descriptions for prokka-1.9.1-testing

[aleimba]

New pull request to include multiple EC_number tags per CDS with version prokka-1.9.1-testing and some bug fixes

For this purpose the database format was adapted to include several EC_numbers separated by a semicolon, e.g.:

K9NBS6 3.5.1.13;3.5.1.14;3.5.1.4~~~aam~~~Acylamidase
MTEQNLHWLSATEMAASVASNNLSPNEIAEAMIQRVDAVNPS...

This is included in the necessary database preparations scripts:

• prokka-biocyc_to_fasta_db
• prokka-genbank_to_fasta_db
• prokka-genpept_to_fasta_db
• prokka-hamap_to_hmm
• prokka-uniprot_to_fasta_db

prokka can handle this format now. As a result a new uniprot database in /db/kingdom/Bacteria/sprot is included.

Additionally, a couple of bug fixes:

• Adapted the usage of prokka-hamap_to_hmm (will concatenate result files to 'HAMAP.hmm') and prokka-uniprot_to_fasta_db (divert stdout to '> sprot')
• /protein description initials are changed to lowercase (as instructed by NCBI), except for 'Rossman' and 'Willebrand', see https://github.com/aleimba/prokka/blob/master/bin/prokka#L1177-1178
• I didn't get the if-condition in line 58 of prokka-uniprot_to_fasta_db. What is 'if (1)' for? https://github.com/Victorian-Bioinformatics-Consortium/prokka/blob/master/bin/prokka-uniprot_to_fasta_db#L58
  I removed the condition.
• Included Shaun's 'Remove stray space from locus_tag' Remove stray space from locus_tag #15 pull request, so there's no conflict between our pull requests

This might warrant a new version, changed it to prokka-1.9.2.-testing. ChangeLog.txt is updated as well.

[aleimba]

Forgot to mention, in prokka /gene and /EC_number feature tags are first removed (if it's a hypothetical protein) and subsequently added, see:
https://github.com/Victorian-Bioinformatics-Consortium/prokka/blob/master/bin/prokka#L915-931
Thus, I changed the order, to first add the tags and afterwards remove if hypothetical, see:
https://github.com/aleimba/prokka/blob/master/bin/prokka#L919-934