def test_structlmm_assoc():
random = RandomState(1)
n = 20
k = 4
y = random.randn(n, 1)
E = random.randn(n, k)
M = ones((n, 1))
x = 1.0 * (random.rand(n, 1) < 0.2)
slmm = StructLMM(y, M, E, W=E)
slmm.fit(verbose=False)
pv = slmm.score_2dof_assoc(x)
assert_allclose([pv], [0.8470039620073695], rtol=1e-5)
def test_structlmm_inter():
random = RandomState(1)
n = 20
k = 4
y = random.randn(n, 1)
E = random.randn(n, k)
M = ones(n)
x = 1.0 * (random.rand(n) < 0.2)
M = stack([M, x], axis=1)
slmm = StructLMM(y, M, E, W=E)
slmm.fit(verbose=False)
pv = slmm.score_2dof_inter(x)
assert_allclose([pv], [0.6781070640353783], rtol=1e-5)
def struct_lmm(geno_df,
pheno,
env,
covs=None,
rhos=None,
no_association_test=False,
no_interaction_test=False):
"""
Utility function to run StructLMM
Parameters
----------
geno_df: (`N`, `S`) pandas.DataFrame
genotype matrix for `N` samples, `S` SNPis
pheno : (`N`, 1) ndarray
phenotype vector
env : (`N`, `K`)
Environmental matrix (indviduals by number of environments)
covs : (`N`, L) ndarray
fixed effect design for covariates `N` samples and `L` covariates.
rhos : list
list of ``rho`` values.
``rho=0`` correspond to no persistent effect (only GxE);
``rho=1`` corresponds to only persitent effect (no GxE);
By default, ``rho=[0, 0.2, 0.4, 0.6, 0.8, 1.]``
no_association_test : bool
if True the association test is not considered.
The default value is False.
no_interaction_test : bool
if True the interaction test is not considered.
The default value is False.
Returns
-------
res : *:class:`pandas.DataFrame`*
contains pv of joint test (if no_association_test is False),
pv of interaction test (if no_interaction_test is False).
"""
#import pdb; pdb.set_trace()
if covs is None:
covs = sp.ones((env.shape[0], 1))
if rhos is None:
rhos = [0, .2, .4, .6, .8, 1.]
if not no_association_test:
# slmm fit null
slmm = StructLMM(pheno, env, W=env, rho_list=rhos)
null = slmm.fit_null(F=covs, verbose=False)
if not no_interaction_test:
# slmm int
slmm_int = StructLMM(pheno, env, W=env, rho_list=[0])
t0 = time.time()
#import pdb; pdb.set_trace()
res = pd.DataFrame(data=[], index=geno_df.columns.values)
pv = sp.zeros(geno_df.shape[1])
pv_int = sp.zeros(geno_df.shape[1])
for snp in xrange(geno_df.shape[1]):
x = geno_df.values[:, [snp]]
if not no_association_test:
# association test
_p, _rho = slmm.score_2_dof(x) #second arg: optimal rho
pv[snp] = _p
if not no_interaction_test:
# interaction test
covs1 = sp.hstack((covs, x))
null = slmm_int.fit_null(F=covs1, verbose=False)
_p, _rho = slmm_int.score_2_dof(x)
pv_int[snp] = _p
# add pvalues to res
if not no_association_test:
res['pv'] = pv
if not no_interaction_test:
res['pv_int'] = pv_int
t = time.time() - t0
#print '%.2f s elapsed' % t
return res
def st_sscan(G, y, E, M=None, tests=None, verbose=True):
"""Mixed-model with genetic effect heterogeneity.
Parameters
----------
pheno : (`N`, 1) ndarray
phenotype data
environments : (`N`, `E`) ndarray
environments data.
covs : (`N`, `D`) ndarray
covariate design matrix.
By default, ``covs`` is a (`N`, `1`) array of ones.
tests : list
Which tests are performed.
Element list values are ``'inter'`` and ``'assoc'``.
By default, only the interaction test is considered.
rhos : list
for the association test, a list of ``rho`` values must be specified.
The choice of ``rho`` affects the statistical power of the test
(for more information see the StructLMM paper).
By default, ``rho=[0, 0.1**2, 0.2**2, 0.3**2, 0.4**2, 0.5**2, 0.5, 1.]``
verbose : (bool, optional):
if True, details such as runtime as displayed.
"""
from struct_lmm import StructLMM
from numpy import zeros, hstack, asarray
from pandas import DataFrame
rhos = [0.0, 0.1**2, 0.2**2, 0.3**2, 0.4**2, 0.5**2, 0.5, 1.0]
with session_block("struct-lmm analysis", disable=not verbose):
with session_line("Normalising input... ", disable=not verbose):
data = conform_dataset(y, M, G=G, K=None)
y = data["y"]
M = data["M"]
G = data["G"]
if tests is None:
tests = ["inter"]
if "inter" in tests:
slmi = StructLMM(asarray(y, float), E, W=E, rho_list=[0])
if "assoc" in tests:
slmm = StructLMM(asarray(y, float), E, W=E, rho_list=rhos)
slmm.fit_null(F=asarray(M, float), verbose=False)
_pvi = zeros(G.shape[1])
_pva = zeros(G.shape[1])
for snp in range(G.shape[1]):
x = asarray(G[:, [snp]], float)
if "inter" in tests:
# interaction test
M1 = hstack((M, x))
slmi.fit_null(F=M1, verbose=False)
_pvi[snp] = slmi.score_2_dof(x)
if "assoc" in tests:
# association test
_pva[snp] = slmm.score_2_dof(x)
data = OrderedDict()
data["pvi"] = _pvi
data["pva"] = _pva
return DataFrame(data)