# CAFE5 on Rackham@UPPMAX
https://github.com/hahnlab/CAFE5

wget https://github.com/hahnlab/CAFE5/archive/refs/tags/v5.0.tar.gz
tar xvf 5.0.tar.gz

# start clean
module purge
module load uppmax

# These 4 lines are needed to compile (firts 2 to run)
module load gcc/12.2.0 blas/3.6.0
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/sw/comp/intel/compilers_and_libraries_2020.4.304/linux/mkl/lib/intel64
export LIBRARY_PATH=$LIBRARY_PATH:/sw/comp/intel/compilers_and_libraries_2020.4.304/linux/mkl/lib/intel64 
export CPATH=$CPATH:/sw/comp/intel/compilers_and_libraries_2020.4.304/linux/mkl/include

cd CAFE5-5.0
autoconf
# Adapt this line if you want to install the binary in specific folder
./configure
make

The binary is in `bin` sub-folder

./bin/cafe5 

Usage: cafe5 [options]

CAFE is a software that provides a statistical foundation for evolutionary inferences about changes in gene family size.
 The program employs a birth and death process to model gene gain and loss across a user-specified phylogenetic tree,
 thus accounting for the species phylogenetic history. The distribution of family sizes generated under this model can
 provide a basis for assessing the significance of the observed family size differences among taxa.

OPTIONS:
  Required Options:
   --infile, -i                 Path to tab delimited gene families file to be analyzed - Required for estimation.
   --tree, -t                   Path to file containing newick formatted tree - Required for estimation.

  Common Options:
   --help, -h                   This help menu.
   --cores, -c                  Number of processing cores to use, requires an integer argument. Default=All available cores.
   --error_model, -e            Run with no file name to estimate the global error model file. This file can be provided
                                    in subsequent runs by providing the path to the Error model file with no spaces (e.g. -eBase_error_model.txt).
   --log_config, -L             Turn on logging, provide name of the configuration file for logging (see example log.config file).
   --n_gamma_cats, -k           Number of gamma rate categories to use. If specified, the Gamma model will be used to run
                                    calculations, otherwise the Base model will be used.
   --output_prefix, -o          Output directory - Name of directory automatically created for output. Default=results.
   --lambda_tree, -y            Path to lambda tree, for use with multiple lambdas.
   --fixed_alpha, -a            Alpha value of the discrete gamma distribution to use in category calculations. If not
                                    specified, the alpha parameter will be estimated by maximum likelihood.
   --fixed_lambda, -l           Value (between 0 and 1) for a single user provided lambda value, otherwise lambda is estimated.
   --fixed_multiple_lambdas, -m Multiple lambda values, comma separated.
   --poisson, -p                Use a Poisson distribution for the root frequency distribution. Without specifying this, a
                                    uniform distribution will be used. A value can be specified -p10 (no space) or --poisson = 10,
                                    otherwise the distribution will be estimated from the gene families.
   --simulate, -s               Simulate families. Optionally provide the number of simulations to generate (-s100 no space, or --simulate = 100)
   --rootdist, -f               Path to root distribution file for simulating datasets.

  Less Common Options:
   --lambda_per_family, -b      Estimate lambda by family (for testing purposes only).
   --pvalue, -P                 P-value to use for determining significance of family size change, Default=0.05.
   --zero_root, -z              Include gene families that don't exist at the root, not recommended.
   --Expansion, -E              Expansion parameter for Nelder-Mead optimizer, Default=2.
   --Iterations, -I             Maximum number of iterations that will be performed in 
                                    lambda search. Default=300 (increase this number if likelihood is still improving when limit is hit).
   --Reflection, -R             Reflection parameter for Nelder-Mead optimizer, Default=1.

## `patchelf`
If you want to remove the module and `LD_LIBRARY_PATH` settings, you can "patchelf" the binary, so it knows where to find the libraries on Rackham by running

patchelf --set-rpath /sw/libs/blas/3.6.0/rackham/lib:/sw/comp/gcc/12.2.0_rackham/lib:/sw/comp/gcc/12.2.0_rackham/lib64:/sw/comp/intel//compilers_and_libraries_2020.4.304/linux/mkl/lib/intel64   ./bin/cafe5

## Contacts:
- [Pavlin Mitev](https://katalog.uu.se/profile/?id=N3-1425)
- [UPPMAX](https://www.uppmax.uu.se/)
- [AE@UPPMAX - related documentation](/8sqXISVRRquPDSw9o1DizQ)


###### tags: `UPPMAX`, `RT-270987`, `CAFE5`, `MKL`