Skip to content

Releases: mkirsche/Iris

1.0.5

14 Jul 20:33
Compare
Choose a tag to compare
v1.0.5

Support absolute paths for out_dir

1.0.4

11 Jan 20:08
Compare
Choose a tag to compare
Update version number and rebuild

1.0.3

30 Jun 17:03
Compare
Choose a tag to compare

Adjust return value to 0 when using usage menu

v1.0.2

30 Jun 15:48
Compare
Choose a tag to compare
  • Small updates to output format
  • Added Jar support
  • File renaming to avoid conflicts with upstream packages

v1.0.1

03 Jan 23:43
Compare
Choose a tag to compare

Version 1.0.1

  • Replaced Falconsense consensus module with Racon, improving speed and accuracy
  • Added script for comparing refined variants to a ground truth set
  • Fixed a few bugs and generally increased stability and robustness to different field names
  • Added a number of additional command line flags to enable greater customization
  • Renamed some scripts to reduce naming conflicts with common filenames such as Settings.java
  • Add INFO fields to indicate whether variants have already been processed and/or refined.

Version 1.0

06 Sep 20:35
Compare
Choose a tag to compare

Improvements of IRIS over the original CrossStitch module include:

  • Improved runtime (approximately 8x faster tested on a GIAB dataset with ~13k insertions)
  • Less dependencies on external software (e.g., GNU Parallel)
  • User-friendly help menu which IRIS outputs when run with incomplete or invalid parameters
  • More flexibility in changing the parameters used in different stages of the pipeline
  • Greater transparency in thread usage (entirely controlled through command line flag)
  • Use of Minimap2 as the default aligner, with continued support for ngmlr
  • Collection of external dependencies as submodules which can all be built with a single script if provided binaries are not compatible with a user's machine
  • Simulation tests with accuracy metrics for insertion sequences compared to the ground truth
  • An option to refine deletions
  • Increased error checking and logging to more easily diagnose problems
  • Creation of a tsv file which logs all refinements and edit distance from the original sequences
  • More intelligent choice of which corrected read to extract the refined sequences from
  • Robustness to nearby variants (i.e., checks to ensure that multiple variants are not collapsed into a single call)
  • Ability to resume a failed run and avoid reprocessing variants