program name: MrGBP version: 1.0 Developer: Samaneh Kouchaki (kouchaki.samaneh@gmail.com)
========dependencies Install the following dependencies
eigen3: http://eigen.tuxfamily.org/index.php?title=Main_Page boost: http://www.boost.org/doc/libs/1_62_0/more/getting_started/unix-variants.html methgl: http://mathgl.sourceforge.net/doc_en/Installation.html
for suggest using g++-4.8 or g++-4.9
if you are installing dependencies locally please add the path to the makefile ========instalation guide
(1) run: make code is ready to use
========whtat does the code do: It is for binning metagenomic contigs.
========support platform MrGBP has been tested on both Mac and Linux platforms.
========run There is a number of options:
-fa : the fasta file to be processed by MLBP_BIN
-svdd : the SVD dimension, default: 60
-mlbpn : the MLBP window length, default: 8
-outdir : a directory that all the MLBP_BIN results will be saved including cluster labels and .png figure
-no_clust : the number of clusters for kmeans++, default: 10
-covpm : a file contains average coverage depth; the entries should be seprated by single space
-covpstd : a file contains standard deviation coverage depth; the entries should be seprated by single space
-reps : the numerical mapping including EIIP, Real, Integer, Paired, Atomic, default: Integer
-clust : the clustering method kmeans++ or dbscan: dbscan
-dbep : the epsilon for dbscan, default: 0.02
-dbminpt : the number of minimum neighbouring points for dbscan, default: 8
-save_feat : If the features needs to be saved separately.
-no_dims : bh-tSNE dimensions, defualt: 2.
examples:
- Running with default option on a fasta file indicated by the path and save the results in desired directory
./MrGBP -fa /pathtofile/file.fa -outdir /pathtofile/file.fa
- Check the available options
./MrGBP -help
- run using coverage profile
./MrGBP -fa /pathtofile/file.fa -covpm /pathtofile/cpm -covpstd /pathtofile/cpstd
Codes for kmeans++ and dbsacan have been downloaded and modified to match our code: kmeans++: http://rosettacode.org/wiki/K-means%2B%2B_clustering dbscan: https://github.com/propanoid/DBSCAN
if there is any question or any bug please email me at kouchaki.samaneh@gmail.com