| NEWS | R Documentation |
genepop News
CHANGES UP TO VERSION 1.2.14
Fixed compilation issue with g++ -std=gnu++20.
CHANGES UP TO VERSION 1.2.10
FIXED BUGS
Mantel test result was incorrect when the rank statistic was not used.
A bug had been introduced in version 1.2.2, in the computation of the 'a' statistic of isolation by distance, in the presence of missing genotypes for some combinations of individuals and loci. The effect on the estimated slope of the IBD regression could be disproportionately large, so any result obtained with the incorrect code is suspect even if only a few missing genotypes are missing from the data.
PACKAGE INSTALLATION
The C++11 specification for the C++ code has been dropped.
DOCUMENTATION
Links to new Genepop reference page.
CHANGES UP TO VERSION 1.2.2
NEW FEATURES
Two additional bootstrap variants have been implemented for analyses of isolation by distance, their selection and usage being controlled by new arguments of ibd(), 'bootstrapMethod' and 'bootstrapNsim'. One of the method (BC) is of particular interest for data sets of thousands of loci.
New 'Batch2file' mode of running Genepop, suitable for use in execution environments where the console output is redirected to a file.
Computation of the 'e' statistic of isolation by distance accounts for missing genotypic information in each pair of individual differently from previosu version. The effect is small unless your data are lousy.
CHANGES UP TO VERSION 1.1.4
CORRECTED BUGS
Incorrect final table of multi-locus tests of pairwise differentiation [test_diff(...,pairs = TRUE)] particularly in cases where all individual tests had p-value estimates equal to 0 (per-locus per-pair tests were correctly reported). Several minor display glitches in similar tables have also been fixed (as a result, scientific format for numbers may now appear in such outputs).
CHANGES UP TO VERSION 1.1.3
NEW FEATURES
New convenience function genepopExample() replacing potentially confusing code in the Examples.
CHANGES UP TO VERSION 1.1.2
NEW FEATURES
New keyword 'intra_all_types' for setting "popTypeSelection" allows one to perform a single spatial regression for all pairs of individuals or populations within several types of subpopulations.
The IBD bootstrap no longer writes many temporary files. This removes an exposure to operating system-dependent crashes.
(R package only) Progress bar for some long computations.
New function clean_workdir() to remove easily files written by Genepop (including output files!).
CORRECTED BUGS
Missing pairwise genetic information was no correctly detected in Mantel tests.
Fixed documentation bug referring to PopTypeSelection="inter_all_pairs" instead of the correct PopTypeSelection="inter_all_types".
CHANGES UP TO VERSION 1.0.5
NEW FEATURES
The name extension and accessibility of the genepop input file is now checked within the R code.
Existence of the output file is checked.
More accurate messages when differentiation options are run on an input file containing only one 'pop'.
CORRECTED BUGS
Linkage disequilibrium tests incorrectly indicated 'No data' for tests involving the last locus.
The ibd() last line of output in the <>.IBD file mis-reported an uninteresting computation. Both the computation and the line of output have been removed.
(R package only) Mantel test always used rank correlation statistic.