Filtus

------------------------------
BUG FIXES:
* Column filters in start-up filters didn't work properly in the last version.

* Loading of variant exclusion databases was unnecessarily repeated.

* Memory leak fixed.

-----------------------------
BUG FIXES:
* Fixed bug in the 'Summarize column' function, which couldn't handle certain cases with empty variant files and/or files missing the indicated column.

------------------------------

NEW FEATURES:
* A new feature "Prefilter" replaces the old "Startup filter" in the Input settings dialog. If a prefilter is given, it is applied to each line before processing it (splitting columns a.s.o), making it fast and memory efficient. (In contrast, the old "Startup filter" processed the whole file before applying the filters.)
By using prefilters you can work with files of virtually any size in FILTUS, as long as the prefilter reduces it to a manageable number of variants. Example: Suppose you have a whole-genome variant file, with some annotation indicating the functional type of each variant. A useful prefilter in this case could be: "Keep lines which contain 'exonic OR splice'"

* The de novo detection feature is moved to the Analysis menu, and has more options (including filtering on ALT ratios).

-----------------------------
NEW FEATURES:
* The 'Summarize column' function now gives a sensible summary of numerical columns, including the minimum, maximum and mean of the non-missing entries. If the total number of unique entries is less than 25, the column is treated as categorical.

BUG FIXES:
* Partly rewritten the file loading methods, motivated by a bug (�B) which sometimes caused crash when loading several files with non-default parameters.

* Fixed minor bug in the 'Table' function

OTHER:
* Faster file loading in 'do-nothing' cases, i.e. where no columns are to be split or deleted, and no startup filters are applied.

-----------------------------
NEW FEATURES:
* Export to MERLIN format. The 'File' menu has an entry for exporting selected files to MERLIN format (ped/map/dat).
Note: Pedigrees are not implemented in FILTUS yet. Thus the pedigree part (first 6 columns) of the pedfile must be manually edited afterwards.

BUG FIXES:
* The 'Build database' function: If a variant occurs more than once within a file, only the first occurrence is used.

* Minor fixes and improvements.

------------------------------
NEW FEATURES:
* New option in the Filter menu: Individual filtering. Up to 4 different filter configuration files can be applied to specified individuals. (Experimental, needs more testing.)

* Meta information in output files. By default, a preamble is included in the beginning of all output files, with information on the samples, filters and analysis involved. Can be switched off in the Preferences menu.