PLINK 1.9 and 2.0
Development of the baseline PLINK toolset has resumed and beta testing for the next (1.90) release has now begun.
The new release, by Chris Chang and colleagues, is a complete rewrite of the original code and represents a very significant improvement in overall speed and functionality. Moving forward, these changes should enable PLINK to meet the demands of ever-larger genetic datasets.
For download and documentation: https://www.cog-genomics.org/plink2.
New features include:
- Very large speed improvements across the board. Basic data management operations tend to be at least 5x as fast; speedup factors for more CPU-intensive functions such as --blocks, --fast-epistasis, --genome, --indep-pairwise, and --r2 frequently exceed 100x on multicore systems.
- The new O(sqrt(n))-time implementations of SNP-HWE and Fisher's exact test can easily be adapted for inclusion in other software, and may be of particular interest to other tool developers.
- Improved memory efficiency.
- Conveniences such as direct VCF import/export and improved command-line help.
Questions, comments, and bug reports should now be directed to the plink2-users Google group:
https://groups.google.com/d/forum/plink2-users
The original version of PLINK will continue to be available for download at the original website.
PLINK 1.9 is currently under heavy active development -- users may wish to check certain results against the original PLINK (1.07) and notify the above list of any discrepancies.
|