NEWS

seqCAT 1.8.0

The Python implementation of profile creation has been deprecated
SNV profiles are now stored as data frames, rather than GRanges objects
The 'create_profile' function now now reads profiles into memory, instead of storing them on disk. De-duplication and removal of mitochondrial variants is now also performed at this stage
The 'create_profiles' function now returns a list of data frames
A new function, 'write_profile', can write profiles to disk
The 'read_profile' function now reads profiles without performing any de-duplication or removing mitochondrial chromosomes
The 'compare_profiles', 'compare_many' and 'list_variants' functions now converts input profiles to GRanges internally
Keep the FILTER column in created SNV profiles
Add a check for the creation of zero-variant profiles
Add a check for the existance of the specified input samples
Removal of non-standard chromosomes and variant de-duplication is now optional, and filtration documentation has been extended

Make the variant caller-specific filtering optional (on by default) and rename the relevant parameter names for increased clarity
Add a check to look for gVCF files as input, including the existance of NON_REF alleles (common for gVCF files)
Add checks to see if input VCFs correctly contain DP, AD and GT data

Add convenience functions for creating and reading multiple SNV profiles
Add functionality for reading general COSMIC mutational data, not just cell line mutational data

Add functionality for analysing VCF files containing unannotated variants
Add functionality for listing non-overlapping variants between profiles
Mitochondrial variants can now be optionally skipped when reading SNV profiles in the 'read_variants' function
Add the 'list_variants' function for listing the genotypes of user-specified variants in each provided SNV profile
Add the 'plot_variant_list' function for plotting a genotype grid for each variant output by the 'list_variants' function

Fix a multi-sample VCF profile creation issue (python only)
Reading zero-variant profiles now properly returns a GRanges object with a dummy-variant profile containing the sample name
Enable the 'plot_impacts' function to properly analyse multi-impact SNVs
Fix reading of SNV profiles containing single-quoted strings