Compiling

GNU g++-7 must be installed. In order to compile Look4TRs, run the following commands

cd Look4TRs

mkdir bin

cd bin

cmake ..

make

Splitting fasta files

Look4TRs is designed to read in multiple fasta files where each file represents a differnt chromosome. However, the fasta file format specifies that an entire genome can be placed into a single file. To remedy this, we have provided a file under src/ called divideGenomeIntoChromosomesEfficient.pl. divideGenomeIntoChromosomesEfficient.pl will divide the FASTA file into multiple files based on their headers. To call this, use the following command

perl divideGenomeIntoChromosomesEfficient.pl Input.fa OutputDirectory

Then, call Look4TRs using OutputDirectory as the input directory from which to read .fa files (using the --adr parameter)

Running

To run Look4TRs on a genome, you'll need two folders consisting of the input directory and output directory. The input directory should contain .fa files. Run the following command to obtain microsatellites.

Look4TRs --adr Input --out Output --default

In order to receive additional help, run one of the following commands

Look4TRs

Look4TRs --help

Look4TRs -h

Parameters

REQUIRED OPTIONS

--adr

Takes the address of the directory of the sequences to be
scanned. Each file must be a FASTA file and end with .fa

--out

Takes the address of the output prediction directory.

NOTE: After supplying the required options, either provide a single default option, or provide all of the parameter options!

DEFAULT OPTIONS

--default

Runs the tool with the default parameters.

PARAMETER OPTIONS

--trn

Takes the address of the directory of the training files.

--motif-location-file <file.bed>

This optional will make Look4TRs train using repeats that are specified by a bed file.
Doing this will mean that minm and maxm will not have to be initialized.
WARNING: This file should correspond with the file used for training Look4TRs or the file indicated by --motif-fa-file!

--motif-fa-file <file.fa>

This will determine the fa file used to generate repeats for training.
If ommited, the training file (either specifed by the --trn flag or the largest one in the --adr directory) will be used.
If included, a bed file (initialized by --motif-loc-file) must be included with it.

--min

The minimum k-mer length analyzed (must be less than or equal
to the maximum k-mer length).
This is used by the scoring module.
Default is 4.

--max

The maximum k-mer length analyzed (must be greater than or equal
to the minimum k-mer length).
This is used by the scoring module.
Default is 6.

--int

The minimum allowed region size for predictions.
(Must be greater than 0)
Default is 20.

--win

The number of multiples of --int to use. (must be positive)
This will allow Look4TRs to auto-callibrate the scoring module.
Default is 4.

--hmm-states-upper

the upper bound for the number of hmm states used
(must be positive, even, and higher than or equal to --hmm-states-lower)
Default is 12.

--hmm-states-lower integer

The lower bound for the number of HMM states used
(must be postive, even, and lower than or equal to --hmm-states-upper)
Default is 6.

--seg

Determines the size of the training chromosome (must be positive)
Default is 3,000,000

--mimn

The minimum size of the motifs used to generate the training Chromosome.
(must be at least 1 and less than or equal to maxm)
Default is 1.

--maxm

The maximum size of the motifs used to generate the training Chromosome.
(must be at least 1 and bigger than or equal to minm)
Default is 10.

--ord

The markov order used to generate the training chromosome.
(must not be negative)
Default is 0.

--mtf <0|1>

 Enable ('1') or disable ('0') the motif discovery feature.
 Default is 1.

--prn

 The number of times that the motif analyzer will train on motifs. Must be positive.
 Default is 2000.

--idn

The minimum allowed identity score between the exact repeat and a candidate (Must be 0 or greater).
Default is 50.

--save-data

A directory where the -hmm, --glm, and --chmp files will be stored.

--hmm

The hmm file that is produced from a previous run that used --save-data.
This HMM will be loaded, instead of training one.
Note: If --hmm, --glm, or --chmp is used, all of them must be used!

--glm

The glm file that is produced from a previous run that used --save-data.
This GLM will be loaded, instead of training one.
Note: If --hmm, --glm, or --chmp is used, all of them must be used!

--chmp

The file that is produced from a previous run that used --save-data.
This will be used for parameters used from training the HMM and GLM.
Note: If --hmm, --glm, or --chmp is used, all of them must be used!

--get-testing-data

The directory where the training and testing chromosomes will be stored.
These chromosomes are the ones used to auto-callibrate Look4TRs.
This will also store the locations of the tandem repeats

--hmm-states

The numeber of HMM states (must be positive and even)

--thr

Number of threads.
Default is the total number of cores detect on the user's computer.