Interaction-analysis

Lasso

Input file:

1. Predictor (x) file: a csv file, samples are rows and variables are columns.
2. Independent (y) file: a csv file, samples are rows and variables are columns.

Script usage:

./lasso.R \
-s . \ # Path to the script
--input_predictor_file /INPUT/X/FILE/NAME \ # Name of  the input predictor(x) file (csv file,column:taxa name, row:sample)
--input_independent_file /INPUT/Y/FILE/NAME \ # Name of the input independent(y) file (csv file,column:gene name, row:sample)
-n 6 \ # Number of threads
-o  /OUTPUT/PATH \ # Path to the output
-x taxa \ # Name of predictor variable
-y gene # Name of independent variable

./lasso.R --help # check for usage information of the script.

Output files:

1. output_path/Rawfile

each file represents for a single independent variable with all the predictors.

column name	explaination
independent variable name	the parameter specified as -x
predictor variable name	the parameter specified as -y
r.sqr	R-squared of the smallest mean squared error model giving by lambda.min during the cross validation process
r.sqr.adj	Adjusted R-squared of the smallest mean squared error model giving by lambda.min during the cross validation process
pval	p value for each predictor varible of the model after desparse lasso estimation analysis
padj	Adjusted p value for each predictor varibale
ci.lower	The lower confidence interval of coeffient of each predictor variable
ci.upper	The upper confidence interval of coeffient of each predictor variable
sigma	The standard error of the model on the whole dataset instead of cross validation
sigma.flag	1 if the number of non-zero coeffient predictor variables is too big (more than n-2 , n is the sample number)
selected	yes if the predictor variable is stably selected by stablity selection

2. Name_lasso_results_FDR_0.05_stable.csv

combine all the independent variable files and adjusted for the p value, only those with FDR < 0.05 and stably choosed rows are retained.

SparseCCA

Input files:

1. Predictor (x) file: a csv file, samples are rows and variables are columns.
2. Independent (y) file: a csv file, samples are rows and variables are columns.

Script usage:

./sparseCCA.R \
--script_path /SCRIPTS/PATH \
--input_predictor_file /INPUT/X/FILE/NAME \
--input_independent_file /INPUT/Y/FILE/NAME \
-n 8 \
-o /OUTPUT/PATH \
--name Control 

Output files:

1. Name/xx_xx.txt
2. Name/components/independent_predictor_component_x.txt