The ProbABEL development team is happy to announce the release of ProbABEL v0.4.5. This release is a bugfix update. People using the Cox Proportional Hazards regression module are strongly advised to upgrade.
(Please click on "read more", below, for more details.)
Information on how to get the source code or binaries can be found on the ProbABEL page
The most important user-visible changes are (note that for users who only use the linear or logistic regression modules, nothing has changed):
- Fix of bug #6041: ProbABEL's Cox module reports too many errors (beta may be infinite, setting beta and se to 'NaN'). Part of the checks were incorrectly implemented and as a result some SNPs were incorrectly thrown out.
- Fix of bug #1266: pacoxph doesn't run without covariates. For a long time (this bug was first reported in 2011) the Cox module needed at least one covariate in order to run. With the introduction of bug #6041 this couldn't be worked around anymore by adding a constant covariate. Now pacoxph behaves correctly in this respect.
- Fix of bug #6054: Not all ProbABEL's short command line options are correctly parsed. The command line options -v (mmscore) and -k (interaction_only) were not correctly parsed if short options were used. No errors would have been displayed if the mandatory argument was missing.
- The regression equation used by pacoxph wasn't always correctly displayed (with e.g. extra + signs). This has been fixed now
- Minor fixes/improvements to the documentation and man-pages
Thanks to all who made this release possible, by providing code patches, by being beta testers, by suggesting enhancements, or otherwise. Especially Anne Grotenhuis and forum user quentin300.