numpy.savetxt(fname=folder+'thinned_S_ccle_ec', X=thinned_S_ccle_ec)
"""
''' Store the mean of the matrices. '''
folder = project_location + 'HMF/drug_sensitivity/bicluster_analysis/matrices_nonneg_5_5/'

E_drugs, E_cell_lines = 'Drugs', 'Cell_lines'
n_gdsc, n_ctrp, n_ccle_ic, n_ccle_ec = 0, 1, 2, 3

exp_F_drugs = HMF.approx_expectation_Ft(E=E_drugs,
                                        burn_in=burn_in,
                                        thinning=thinning)
exp_F_cell_lines = HMF.approx_expectation_Ft(E=E_cell_lines,
                                             burn_in=burn_in,
                                             thinning=thinning)
exp_S_gdsc = HMF.approx_expectation_Sn(n=n_gdsc,
                                       burn_in=burn_in,
                                       thinning=thinning)
exp_S_ctrp = HMF.approx_expectation_Sn(n=n_ctrp,
                                       burn_in=burn_in,
                                       thinning=thinning)
exp_S_ccle_ic = HMF.approx_expectation_Sn(n=n_ccle_ic,
                                          burn_in=burn_in,
                                          thinning=thinning)
exp_S_ccle_ec = HMF.approx_expectation_Sn(n=n_ccle_ec,
                                          burn_in=burn_in,
                                          thinning=thinning)

numpy.savetxt(fname=folder + 'F_drugs', X=exp_F_drugs)
numpy.savetxt(fname=folder + 'F_cell_lines', X=exp_F_cell_lines)
numpy.savetxt(fname=folder + 'S_gdsc', X=exp_S_gdsc)
numpy.savetxt(fname=folder + 'S_ctrp', X=exp_S_ctrp)
Exemple #2
0
HMF = HMF_Gibbs(R, C, D, K, settings, hyperparameters)
HMF.initialise(init)
HMF.run(iterations)
''' Store the mean of the matrices. '''
folder = project_location + 'HMF/methylation/bicluster_analysis/matrices/'

E_drugs, E_cell_lines = 'genes', 'samples'
n_ge, n_pm, n_gm = 0, 1, 2

exp_F_genes = HMF.approx_expectation_Ft(E=E_drugs,
                                        burn_in=burn_in,
                                        thinning=thinning)
exp_F_samples = HMF.approx_expectation_Ft(E=E_cell_lines,
                                          burn_in=burn_in,
                                          thinning=thinning)
exp_S_ge = HMF.approx_expectation_Sn(n=n_ge,
                                     burn_in=burn_in,
                                     thinning=thinning)
exp_S_pm = HMF.approx_expectation_Sn(n=n_pm,
                                     burn_in=burn_in,
                                     thinning=thinning)
exp_S_gm = HMF.approx_expectation_Sn(n=n_gm,
                                     burn_in=burn_in,
                                     thinning=thinning)

numpy.savetxt(fname=folder + 'F_genes', X=exp_F_genes)
numpy.savetxt(fname=folder + 'F_samples', X=exp_F_samples)
numpy.savetxt(fname=folder + 'S_ge', X=exp_S_ge)
numpy.savetxt(fname=folder + 'S_pm', X=exp_S_pm)
numpy.savetxt(fname=folder + 'S_gm', X=exp_S_gm)
Exemple #3
0
def test_approx_expectation():
    iterations = 10
    burn_in = 2
    thinning = 3 # so index 2,5,8 -> m=3,m=6,m=9
    
    E = ['entity0','entity1']
    I = {E[0]:5, E[1]:3}
    K = {E[0]:2, E[1]:4}
    J = [6]
    
    iterations_all_Ft = {
        E[0] : [numpy.ones((I[E[0]],K[E[0]])) * 3*m**2 for m in range(1,10+1)],
        E[1] : [numpy.ones((I[E[1]],K[E[1]])) * 1*m**2 for m in range(1,10+1)]
    }
    iterations_all_lambdat = {
        E[0] : [numpy.ones(K[E[0]]) * 3*m**2 for m in range(1,10+1)],
        E[1] : [numpy.ones(K[E[1]]) * 1*m**2 for m in range(1,10+1)]
    }
    iterations_all_Sn = [[numpy.ones((K[E[0]],K[E[1]])) * 2*m**2 for m in range(1,10+1)]]
    iterations_all_lambdan = [[numpy.ones((K[E[0]],K[E[1]])) * 2*m**2 for m in range(1,10+1)]]
    iterations_all_taun = [[m**2 for m in range(1,10+1)]]
    iterations_all_Sm = [[numpy.ones((K[E[1]],K[E[1]])) * 2*m**2 * 2 for m in range(1,10+1)]]
    iterations_all_lambdam = [[numpy.ones((K[E[1]],K[E[1]])) * 2*m**2 * 2 for m in range(1,10+1)]]
    iterations_all_taum = [[m**2*2 for m in range(1,10+1)]]
    iterations_all_Gl = [[numpy.ones((J[0],K[E[1]])) * 2*m**2 * 3 for m in range(1,10+1)]]
    iterations_all_taul = [[m**2*3 for m in range(1,10+1)]]
    
    expected_exp_F0 = numpy.array([[9.+36.+81. for k in range(0,2)] for i in range(0,5)])
    expected_exp_F1 = numpy.array([[(9.+36.+81.)*(1./3.) for k in range(0,4)] for i in range(0,3)])
    expected_exp_lambda0 = numpy.array([9.+36.+81. for k in range(0,2)])
    expected_exp_lambda1 = numpy.array([(9.+36.+81.)*(1./3.) for k in range(0,4)])
    expected_exp_Sn = numpy.array([[(9.+36.+81.)*(2./3.) for l in range(0,4)] for k in range(0,2)])
    expected_exp_lambdan = numpy.array([[(9.+36.+81.)*(2./3.) for l in range(0,4)] for k in range(0,2)])
    expected_exp_taun = (9.+36.+81.)/3.
    expected_exp_Sm = numpy.array([[(18.+72.+162.)*(2./3.) for l in range(0,4)] for k in range(0,4)])
    expected_exp_lambdam = numpy.array([[(18.+72.+162.)*(2./3.) for l in range(0,4)] for k in range(0,4)])
    expected_exp_taum = (18.+72.+162.)/3.
    expected_exp_Gl = numpy.array([[(27.+108.+243.)*(2./3.) for k in range(0,4)] for j in range(0,6)])
    expected_exp_taul = (27.+108.+243.)/3.
    
    R0, M0 = numpy.ones((I[E[0]],I[E[1]])), numpy.ones((I[E[0]],I[E[1]]))
    C0, M1 = numpy.ones((I[E[1]],I[E[1]])), numpy.ones((I[E[1]],I[E[1]]))
    D0, M2 = numpy.ones((I[E[1]],J[0])), numpy.ones((I[E[1]],J[0]))
    R, C, D = [(R0,M0,E[0],E[1],1.)], [(C0,M1,E[1],1.)], [(D0,M2,E[1],1.)]
    
    alphatau, betatau = 1., 2.
    alpha0, beta0 = 6., 7.
    lambdaF, lambdaG = 3., 8.
    lambdaSn, lambdaSm = 4., 5.
    priors = { 'alpha0':alpha0, 'beta0':beta0, 'alphatau':alphatau, 'betatau':betatau, 
               'lambdaF':lambdaF, 'lambdaG':lambdaG, 'lambdaSn':lambdaSn, 'lambdaSm':lambdaSm }
    settings = { 'priorF' : 'exponential', 'priorG' : 'normal', 'priorSn' : 'normal', 'priorSm' : 'normal', 
                 'orderF' : 'columns', 'orderG' : 'rows', 'orderSn' : 'rows', 'orderSm' : 'rows',
                 'ARD' : True, 'element_sparsity': True }    
    
    HMF = HMF_Gibbs(R,C,D,K,settings,priors)
    HMF.iterations = iterations
    HMF.iterations_all_Ft = iterations_all_Ft
    HMF.iterations_all_lambdat = iterations_all_lambdat
    HMF.iterations_all_Sn = iterations_all_Sn
    HMF.iterations_all_lambdan = iterations_all_lambdan
    HMF.iterations_all_taun = iterations_all_taun
    HMF.iterations_all_Sm = iterations_all_Sm
    HMF.iterations_all_lambdam = iterations_all_lambdam
    HMF.iterations_all_taum = iterations_all_taum
    HMF.iterations_all_Gl = iterations_all_Gl
    HMF.iterations_all_taul = iterations_all_taul
    
    exp_F0 = HMF.approx_expectation_Ft(E[0],burn_in,thinning)
    exp_F1 = HMF.approx_expectation_Ft(E[1],burn_in,thinning)
    exp_lambda0 = HMF.approx_expectation_lambdat(E[0],burn_in,thinning)
    exp_lambda1 = HMF.approx_expectation_lambdat(E[1],burn_in,thinning)
    exp_Sn = HMF.approx_expectation_Sn(0,burn_in,thinning)
    exp_lambdan = HMF.approx_expectation_lambdan(0,burn_in,thinning)
    exp_taun = HMF.approx_expectation_taun(0,burn_in,thinning)
    exp_Sm = HMF.approx_expectation_Sm(0,burn_in,thinning)
    exp_lambdam = HMF.approx_expectation_lambdam(0,burn_in,thinning)
    exp_taum = HMF.approx_expectation_taum(0,burn_in,thinning)
    exp_Gl = HMF.approx_expectation_Gl(0,burn_in,thinning)
    exp_taul = HMF.approx_expectation_taul(0,burn_in,thinning)
    
    assert numpy.array_equal(expected_exp_F0,exp_F0)
    assert numpy.array_equal(expected_exp_F1,exp_F1)
    assert numpy.array_equal(expected_exp_lambda0,exp_lambda0)
    assert numpy.array_equal(expected_exp_lambda1,exp_lambda1)
    assert numpy.array_equal(expected_exp_Sn,exp_Sn)
    assert numpy.array_equal(expected_exp_lambdan,exp_lambdan)
    assert expected_exp_taun == exp_taun
    assert numpy.array_equal(expected_exp_Sm,exp_Sm)
    assert numpy.array_equal(expected_exp_lambdam,exp_lambdam)
    assert expected_exp_taum == exp_taum
    assert numpy.array_equal(expected_exp_Gl,exp_Gl)
    assert expected_exp_taul == exp_taul