get_maps

Takes a list of Kegg KOs as input in a file and return a list of Kegg pathways with the number of KO including in each pathway.

get_maps.py -i FILE [-o FOLDER] [-l FOLDER]

-i: file with KOs
-o: output folder [.]
-l: log folder [.]

build_pathways_matrix

The script takes as input pathway counts file generated with get_maps.py and create a count matrix.

build_pathways_matrix.py -f FOLDER [-o FILE]

-f FOLDER folder containing pathways count files
-o FILE output matrix name [KO_matrix.csv]

rf_functions

Utility functions for random forests using the randomForest package in R.

rf_overfitting_test <- function(table, target, mtry, ntree, index_split, iteration, type, maxnode = NULL, print=TRUE)

Compute the difference in Mean Squared Error between the model and predictions. Usefull for determining if a model is overfitting or not.

best_rf <- function(table, target, mintree, maxtree, step)

Find the best combination of mtry and ntree values in order to obtain the best model.

remove_least_imp <- function(table, model, proportion)

Only keep most important predictors. Usefull for decomplexing model, lowering overfitting and calculation time.

gbk_commands

Different functions to use with similar Genbank files. Useful for sequence vizualisation with tools like Clinker.

gbk_commands.py find -i [FOLDER]

Find genes common to all Genbank files. Duplicated genes are ignored.

-i: Folder with all Genbank files

gbk_commands.py order [-h] -i FOLDER -o FOLDER -t STR

Re-order Genbank files considering a gene in common as origin.

-i: Folder with all Genbank files -o: Output folder for all reordered Genbank files -t: Target gene considered as origin

gbk_commands.py extract [-h] -i FOLDER -o FOLDER -t STR -r NUM

Extract region surrounding a specified gene.

-i: Folder with all Genbank files -o: Output folder for all extracted regions -t: Target gene -r: Sequence length to extract before and after the target