Changelogs » Denovonear



- include full intronic region when getting mutation rates
- asynchonously perform ensembl requests
- finding closest exon now much faster, which speeds up many other steps
- roughed out basic consequence determination for protein coding transcripts


fixed bug when retrieving cached data for genes without current HGNC symbols in ensembl


fixed occasional sqlite3 error with concurrent denovonear runs


Shifted to a `denovonear` entry point for CLI purposes. Now can install and run one of:
- `denovonear cluster` (to test for clustering)
- `denovonear transcripts` (to identify transcripts containing de novos)
- `denovonear rates` (to calculate mutation rates per gene)


- Fixed bug when merging transcripts from one gene, which affected a few genes throughout the genome
- Cleaned up loading genes
- Shifted additional functions into package
- removed deprecated IDs option


- swapped clustering function inside package, now can be called from package.
- fixed occasional error when getting union of transcripts, when transcripts have
had their CDS coordinates corrected to make a complete transcript.
- separated cpp code into separate directory
- various fixes to allow for uploading to pypi for pip install
- fixed abs include and abs -> std::abs in cpp code for installs on macos via clang


- fixed merging transcripts when transcripts do not overlap
- use auto-based for-each loops in c++
- better random initialisation
- reduced merge-sort overhead in genes with extreme p-values
- allow for combining WeightedChoice objects (and underlying c++ classes)
- can use genomic coordinates in sites-specific rates


- fixed bug in sampling where genes with p-values < 1e-6 were skewed to be even more significant


updated version number in python package


- can now use larger sequence contexts, such as 7-mers, instead of the
trinucleotide based rates


converted Transcript and SiteRates classes to cython over c++ classes


- refactored how coordinates on + and - strands are handled
- include ref and alt alleles in choices, so we can extract the alleles,
rate and position for a given consequence choice, which allows us to
also use other prediction tools for choices.
- improved unit tests throughout