See #2598.

The regexp and a J matching the regexp are aligned
with sequences that don't match the regexp

See #2133

See https://www.ncbi.nlm.nih.gov/news/09-17-2014-simple-FASTA-headers-genomes-FTP/
Fixes #2173.

See #2132.

Thanks to landscape.io.

Mikaël Salson authored Sep 22, 2016 and Mathieu Giraud committed Sep 29, 2016

The Federal government asked them to do so
<https://www.ncbi.nlm.nih.gov/home/develop/https-guidance.shtml>.

Yeah… encryption is good :)

There are currently no gapped J sequences distributed by IMGT.
We create these sequences aligned on the Phe118/Trp118.

This completes 39bddf7a.

We don't know yet the meaning of this letter. Depending on this answer
we may add some processing

We do not want to mix IGHD (IgH D genes) and IGHC=D (IgD constant region).

The syntax was not working: split the test.

Therefore a sequence can go in several files

In that case the sequence will be ignored

We cannot have a dict of dict: the same gene name can be shared with several entries
so we put an additional list, where the different entries for the same gene will be put
under the same list.

We can have several entries with the same gene name (but the values would differ).
It is not a good idea to store it as a dict.

Thanks to landscape.io

The script allows to retrieve downstream or upstream sequences of some genes whose name (or regex)
are provided. This allows to match more easily DH-JH sequences for instance, DD2-DD3 or also
J genes that are quite short.

Added parenthesis to print for python3 compatibility on line 8 of germline/buildBrowserGermline.py and lines 17 and 27 of germline/split-from-imgt.py.

IMGT considers some sequences as belonging to several germlines. This is the
case for example for TRAV../DV[45678] sequences that are shared by TRAV and
TRDV sequences. Therefore they should be put in both germlines.

For DD2/DD3 recombinations we need respectively upstream and downstream sequences of the two genes.
They are not included in IMGT, therefore we retrieve them from NCBI

Dd2 and Dd3 may play interesting roles.

	* New selection of representative read (core/read_chooser.cpp)
	* Faster spaced seed computation (core/tools.cpp)
	* New unit tests
	* Bugs closed

	* First public release