LIGAP The LIGAP method is widely applicable to quantify differential time-course dynamics of many types of datasets and generalizes to any number of conditions usage: result = ligap(time_vector, data_vector, labels) where: result is the result vector, probabilities has the posterior probabilities for the models specified by models variable (see below), profiles has the means and standard deviations of the fitted GPs per condition (evaluation time points are given) time_vector is a cell vector of length N (N conditions) containing the measurement time points data_vector is a cell vector of length N (N conditions) containing the corresponding measurements (log2 transformed for microarrays) for the time points given in time_vector, order should be the same labels is a cell vector of length N holding the labels of the experimental conditions description of the model notation: {cond1cond2cond3}: all conditions are similar {cond1cond2}{cond3}: cond3 is different cond1 and cond2 (which are similar) {cond1cond3}{cond2}: cond2 is different cond1 and cond3 (which are similar) {cond1}{cond2cond3}: cond1 is different cond2 and cond3 (which are similar) {cond1}{cond2}{cond3}: all conditions are different examples: two conditions, three time points (1,2,3) and three replicates: result = ligap({[1,1,1,2,2,2,3,3,3]',[1,1,1,2,2,2,3,3,3]'}, {rand(9,1),rand(9,1)},{'cond1','cond2'}) three conditions, three time points (1,2,3), and varying number of replicates result = ligap({[1,1,1,2,2,2,3,3,3]',[1,1,2,2,3,3]', [1,1,1,2,3,3,3]'},{rand(9,1),rand(6,1), rand(7,1)},{'cond1','cond2','cond3'}) The GPML Matlab code is written by Carl Edward Rasmussen. The partition code is written by Matt Fig. Author: Tarmo Äijö Updated: 03/05/2013