R package: lfmm (updated December, 2017)
New latent factor algorithms are available in functions of the R package lfmm (recommended for large data sets). For documentation and examples, see lfmm (github devel version).
R version: 1.5 (updated May, 2017)
The LFMM software is available in a function of the R package LEA (recommended use). The R package contains documentation files and examples. See the LEA website.
Version: 1.5 (updated May, 2017)
The LFMM software is available in a command line version (CL) for Linux and Mac Operating System, and with a Graphical User Interface (GUI) for Windows, Linux and Mac Operating System (OS). The package contains documentation files, perl and R scripts and a tutorial example. LFMM does not accept missing genotypes. Use a genotype imputation method before fitting models. See the R function LEA::impute() for an imputation method.
We recommend using the R package LEA 2.0.0 (version (CL_v1.5)) or the R package lfmm (devel version).
- For Linux and Mac: LFMM_CL_v1.5.zip
- For all OS: LEA
Graphical User Interface version (GUI) (v 1.2):
- Linux 64bits: LFMM_GUI_Linux64_v1.2.zip
- Linux 32bits: LFMM_GUI_Linux32_v1.2.zip
- Mac 64bits: LFMM_GUI_Mac64_v1.2.zip
- Windows 64bits: LFMM_GUI_Win64_v1.2.zip
- R lfmm: New algorithms. Faster and sometimes better than the Bayesian boostrap version.
- 1.5: Missing data imputation bug fixed. Replaced by naive imputation. See LEA for recommended usage (imputation of missing data).
- 1.4: Major bug fixed.
- 1.3: Examples showing how to control the FDR, C programs for data conversion, PCA and Tracy-Widom tests. A few bugs fixed. Code factorization.
- 1.2: Bug fixed on the transformation of z-scores in p-values, Windows version not multi-threaded, Windows GUI a bit faster than the previous Windows GUI.
- 1.1: First main version
