Instruction

If you finished to compile 'GAIT', there will be an execution file GAIT in the same folder. The GAIT, input files must be in the same folder.
In terminal, GAIT is executed like the following : ./GAIT inputfile_1 inputfile_2 (OPTION)

The inputfile_1 must follow this form. 0 means 'censored’ and 1 means 'not censored’ in second, forth column.

inputfile_1 form

time_of_event1	is_event1_censored	time_of_event2	is_event2_censored

inputfile_2 form

gene name	expression of sample_1	expression of sample_2   ...	expression of sample_N

compile example

./GAIT input.surv.txt input.expr.txt

options list

   -o filename    : output density file (default screen out)
   -N integer     : the number of T sample (default 1000)
   -D integer     : the number of division (default 11)
   -s b-e,b-e,... : region of interests (default data range)
   -p filename    : using prior density in filename (default not using prior)
   -d integer     : maximum depth of OPT (default 10)
   -n float       : minimum num of effective points per region (default 1)
   -i integer     : maximum number of iterations (default 10)
   -e float       : criteria to stop iterations (default 0.001)
   -r float       : rho (default 0.5)
   -a float       : alpha (default 0.5)
   -P filename    : output phi table (default none)
   -B             : non-boundary mode (default boundary mode)
   -O             : perform OPT once (default iteration)
   -I             : partitioning at integer points (default FALSE)
   -M             : mixed mode on (default off)
   -v             : verb mode
   -h             : show this message

How to run (R version)

Install R package

install.packages("./GAIT.tar.gz")

Example

library(GAIT)
load("./Sample_GAIT.RData");
p.value = GAIT(SurvD = i.s, ExpD = i.e);

We used a nonparametric multivariate survival estimater to solve multivariate censored data. Among options, (-o filename) option makes density of event_1 and event_2 to be generated.

Instruction

How to run (C++ version)

How to run (R version)