Filtus

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

NEW FEATURES:
* Improved de novo calling for mutliallelic variants.
* New boolean parameter "GTonly" in DeNovoComputer.analyze(), indicating whether the GT field alone should decide genotype. If GTonly=False potential false negative variants (based on AD field) will be added.
Note: All variants detected with GTonly=True will also be detected with GTonly=False.

------------------------------
NEW FEATURES:
* Autozygosity mapping, based on a hidden Markov model developed by Kristina S. Gj�tterud. Details and explanations to follow.

* Gene and variant filters: Multiple file names are now allowed in these fields. Separate by semicolon.

* Added the possibility of copying the contents of a single column (in the main text window). Available by right-clicking on the column header.

BUG FIXES:
* Minor cleanups

------------------------------
NEW FEATURES:
* Several changes in the autozygosity mapping dialog (this is still work in progress). The user can now specify a genetic map (e.g. the one included in the 'example_files' folder). If not, the uniform map with 1Mb = 1cM is used everywhere.

* When sorting numerical columns, non-numerical entries (typically missing values) are now always placed last, independent of ascending/descending order.

BUG FIXES:
* The 'de novo' algorithm crashed when encountering missing PL values.

* The previous version crashed on Mac/Linux machines without 'numpy' and 'matplotlib' installed. (These libraries are required only for making plots; everything else should work as normal without them.)

* Fixed a bug in the database 'Search' method.

-----------------------------
NEW FEATURES:
* Added the possibility of truncating columns in the main display. The default maximum width is set to 50 characters, but can be can be changed in the Preferences menu. If set to 0, no truncating is done.
NOTE: Truncation of wide columns does not affect the output of "Save main window content".
NOTE: When working with files with very long lines (e.g. many thousand characters wide) truncation is highly recommended.

-----------------------------
NEW FEATURES:
* Improved region filtering: Regions can now we written directly in the 'Restrict to regions' field, in the format chr:from-to. Use semicolon as separator if more than one region. For example, the region filter 1:0-100; X:200-300
will extract all variants on chromosome between positions 0 and 100, and on the X chromosome between pos 200 and 300.

-----------------------------
NEW FEATURES:
* New filter element: 'Restrict to regions'. This requires a separate text file with one row for each region, and three tab (or space) separated columns: chromosome, start, end. NB: Chromosome numbers must be written as in the variant files.

* New column operators 'starts with' and 'does not start with'.

* When loading a file, a warning is displayed if there are no homozygous variants. (This may happen if a wrong symbol of homozygosity is given in the input settings dialog.)

* More informative output of 'Pairwise sharing'. In addition to variant sharing in absolute numbers, I've added a second matrix showing the results in percentages: The entry (i,j) is the percentage of variants in sample 'i' that are also found in sample 'j'. Note that this is in general not a symmetrical matrix.

* 'Build database' now offers two alternatives: Simple or extended format. The simple format is as before, with five columns (chromosome, position, samples total, heterozygous, homozygous). The extended format has one column per sample, with entries 0 (not present), 1 (heterozygous), 2 (homozygous).
Note: Legal/ethical issues may conflict with using/distributing databases of human data in extended format.

* New function 'Add to database', which adds the variants of the selected files to an existing database. The function recognizes if the database is in simple or extended format. The updated database is sorted and saved in a separate file with a suffix '_new'. (E.g. if the existing file was named spam.txt, the new file will be spam_new.txt.

* The database functions deal differently with repeated variant positions within a single sample. If several variants have the same position, they are counted only once: As homozygous if any of them are, otherwise heterozygous.

BUG FIXES:
* Gene lengths (and therefore p-values in 'Gene sharing' analysis) were incorrectly handled in some cases where the 'Exclude genes' filter was switched on and then off.

* Right click menus didn't show on some Mac systems. Fixed now I think.

OTHER:
* Improved and tightened layout of the filter box.

* Changed to mono-spaced font in all entry fields. (Same font but one point larger than in main text field. Subject to change.)

* In 'Preferences': Added possibility to change the general font size.