Filtus

------------------------------
NEW FEATURES:
* VCF files: If the file contains descriptions of fields in the FILTER, INFO and FORMAT columns, these descriptions are now read by FILTUS and available when the file content is displayed in the main window. Right click on a column header and choose "Description". If the "Description" item is inactive or not present, the column does not have a description.

OTHER:
* Fixed font sizes in several menus.

BUG FIXES:
* Several minor bug fixes.

------------------------------
NEW FEATURES:
* Variant view: By right clicking on a single variant, one can open a separate window showing unfiltered details for this variant in all loaded samples.

* The output of de novo detection has three new columns (after the posterior probability column): The percentage of reads with the ALT allele for the child, father and mother respectively. For a true de novo these are expected to be not too far from 50%, 0% and 0%. (The AD field must be present in the format/genotype columns for this to work.)

BUG FIXES:
* Meta information was not included in "Save main window content" after "Pairwise sharing" analysis.

* Chromosome names with 'chr'-prefixes (e.g. 'chr1' and 'chrX') should work everywhere, including in the gender estimation plot. I think this has been ok for a while, but somehow didn't make it into the changelog.

------------------------------
NEW FEATURES:
* QC plots: The main new feature in this version is a plotting functionality. This imports matplotlib and numpy, making the windows version of FILTUS quite a bit larger. (Note: For Mac and Linux, these libraries must be installed on the system for the plotting to work.) Several types of QC-motivated and exploratory plots are offered. To get a quick overview of a batch of exome sequences it is useful to plot heterozygosity levels, private variants and gender estimation. These can also be compared with an inhouse database of other similar samples. In addition the user can produce histograms or scatter plots of numerical columns in the files.

-----------------------------
NEW FEATURES:
* In the file loading dialog, users can now specify the code for homozygous variants. (In earlier versions this was hard-coded to '1/1' as in VCF format files.) Variants with any other entry in the genotype column are interpreted as heterozygous.

* All entry-fields which expect file numbers (e.g. the Cases and Controls in 'Gene sharing' analysis) now accept inputs of the form 'ID <id1>, <id2>, ...', where each of the words <id1>, <id2>, ... uniquely identifies a file name. Furthermore, entries of the form 'NOT ID <id1>, ...' selects all files except those identified by <id1>, ... .
Example: If the loaded files are 'C:\myfoo.csv', 'C:\foo2.csv', 'C:\bar.csv' and 'C:\bar2.csv', the following entries will all select files 1, 2 and 3:
'1, 2, 3'
'1-3'
'NOT 4'
'ID foo.csv, foo2.csv, bar.csv'
'ID my, o2, r.'
'NOT ID bar2'

BUG FIXES:
* Fixed bug in 'Gene sharing family' with compound heterozygous model: Sometimes the results would include genes with a single heterozygous variant.

* Improved handling of non-variant files

* Improved error handling to avoid crashes with bad file types

* Fixed an error in the output of the 'Table' function in 'Variant sharing'

OTHER:
* Substantially faster filtering in many cases, due to the internal order in which the filters are performed

------------------------------
NEW FEATURES:
* De novo analysis for trios. This analysis requires a single variant file containing genotype data on the child and both parents. The file does not need to be in VCF format, but the genotype columns must be in VCF style, e.g. a FORMAT column followed by the individual genotype columns. The GT and PL fields must be present. When loading the file, make sure to split the FORMAT column, and to check the "Keep 0/0" box.

The algorithm identifies variants where the child is called heterozygous and both parents homozygous for the REF allele. The genotype likelihoods in the PL column are then combined with user specified info on allele frequencies and prior mutation rate, in a bayesian computation giving a posterior probability for each variant of being de novo.

OTHER:
* Improved handling of variant files in VCF and VCF-like formats (e.g. Annovar output files with original VCF columns attached after the annotation columns.) When a FORMAT column with corresponding genotype column(s) is recognized, the genotype column is automatically set to "VCF format" and interpreted as such.

BUG FIXES:
* Scroll bars didn't show properly in some situations