def setUp(self):
self.standardcode = substitution_model.Codon(model_gaps=True,
gc=1,
mprob_model='tuple')
self.mitocode = substitution_model.Codon(model_gaps=False,
gc=2,
mprob_model='tuple')
def test_binned_partition(self):
submod = substitution_model.Codon(
predicates={'kappa': 'transition', 'omega': 'replacement'},
ordered_param='rate', partitioned_params='omega',
distribution='free', mprob_model='tuple')
lf = self._makeLikelihoodFunction(submod, bins=3)
values = list(lf.getParamValueDict(['bin'])['omega_factor'].values())
self.assertEqual(round(sum(values) / len(values), 6), 1.0)
self.assertEqual(len(values), 3)
def test_scaling(self):
"""Testing scaling calculations using Dn and Ds as an example."""
model = substitution_model.Codon(
do_scaling=True, model_gaps=False, recode_gaps=True,
predicates = {
'k': trans,
'r': replacement},
motif_probs={
'TAT': 0.0088813702685557206, 'TGT': 0.020511736096426307,
'TCT': 0.024529498836963416, 'TTT': 0.019454430112074435,
'TGC': 0.0010573059843518714, 'TGG': 0.0042292239374074857,
'TAC': 0.002326073165574117, 'TTC': 0.0086699090716853451,
'TCG': 0.0010573059843518714, 'TTA': 0.020723197293296681,
'TTG': 0.01036159864664834, 'TCC': 0.0082469866779445976,
'TCA': 0.022414886868259674, 'GCA': 0.015648128568407697,
'GTA': 0.014590822584055826, 'GCC': 0.0095157538591668436,
'GTC': 0.0063438359061112285, 'GCG': 0.0016916895749629942,
'GTG': 0.0067667582998519769, 'CAA': 0.018185662930852189,
'GTT': 0.021569042080778176, 'GCT': 0.014167900190315077,
'ACC': 0.0042292239374074857, 'GGT': 0.014167900190315077,
'CGA': 0.0012687671812222456, 'CGC': 0.0010573059843518714,
'GAT': 0.030238951152463524, 'AAG': 0.034891097483611758,
'CGG': 0.002326073165574117, 'ACT': 0.028758722774370905,
'GGG': 0.0071896806935927262, 'GGA': 0.016282512159018821,
'GGC': 0.0090928314654260944, 'GAG': 0.031296257136815393,
'AAA': 0.05476844998942694, 'GAC': 0.011207443434129837,
'CGT': 0.0033833791499259885, 'GAA': 0.076337492070205112,
'CTT': 0.010573059843518714, 'ATG': 0.012687671812222457,
'ACA': 0.021991964474518927, 'ACG': 0.00084584478748149711,
'ATC': 0.0076126030873334746, 'AAC': 0.022837809262000422,
'ATA': 0.017762740537111441, 'AGG': 0.013533516599703954,
'CCT': 0.025586804821315288, 'AGC': 0.029393106364982026,
'AGA': 0.021991964474518927, 'CAT': 0.021357580883907802,
'AAT': 0.05772890674561218, 'ATT': 0.019031507718333687,
'CTG': 0.012899133009092831, 'CTA': 0.013744977796574329,
'CTC': 0.0078240642842038483, 'CAC': 0.0050750687248889825,
'CCG': 0.00021146119687037428, 'AGT': 0.03742863184605625,
'CAG': 0.024106576443222668, 'CCA': 0.021357580883907802,
'CCC': 0.0069782194967223515},
scales = {'dN': replacement, 'dS': ~replacement},
mprob_model = 'tuple',
)
length = 0.1115
a = self._getScaledLengths(model,
{'k': 3.6491, 'r': 0.6317, 'length': length})
b = self._getScaledLengths(model,
{'k': 3.6491, 'r': 1.0, 'length': length})
dN = length * a['dN'] / (3.0 * b['dN'])
dS = length * a['dS'] / (3.0 * b['dS'])
# following are results from PAML
self.assertEqual('%.4f' % dN, '0.0325')
self.assertEqual('%.4f' % dS ,'0.0514')
def test_complex_binned_partition(self):
submod = substitution_model.Codon(
predicates={'kappa': 'transition', 'omega': 'replacement'},
ordered_param='kappa', partitioned_params=['omega'],
mprob_model='tuple')
lf = self._makeLikelihoodFunction(submod,
bins=['slow', 'fast'])
lf.setParamRule('kappa', value=1.0, is_constant=True)
lf.setParamRule('kappa', edge="Human", init=1.0, is_constant=False)
values = list(lf.getParamValueDict(['bin'])['kappa_factor'].values())
self.assertEqual(round(sum(values) / len(values), 6), 1.0)
self.assertEqual(len(values), 2)
def test_codon(self):
"""test a three taxa codon model."""
submod = substitution_model.Codon(
equal_motif_probs=True,
do_scaling=False,
motif_probs=None,
predicates={'kappa': 'transition', 'omega': 'replacement'},
mprob_model='tuple')
likelihood_function = self._makeLikelihoodFunction(submod)
likelihood_function.setParamRule('omega', value=0.5, is_constant=True)
evolve_lnL = likelihood_function.getLogLikelihood()
self.assertFloatEqual(evolve_lnL, -80.67069614541883)
def MG94HKY(**kw):
"""Muse and Gaut 1994 codon substitution model, HKY variant (with kappa,
the ratio of transitions to transversions)
see, Muse and Gaut, 1994, Mol Biol Evol, 11, 715-24"""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'MG94',
predicates = [_kappa, _omega],
mprob_model='monomer',
**kw)
def CNFHKY(**kw):
"""Conditional nucleotide frequency codon substitution model, HKY variant
(with kappa, the ratio of transitions to transversions)
See Yap, Lindsay, Easteal and Huttley, Mol Biol Evol, In press."""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'CNFHKY',
predicates = [_kappa, _omega],
mprob_model='conditional',
**kw)
def CNFGTR(**kw):
"""Conditional nucleotide frequency codon substitution model, GTR variant
(with params analagous to the nucleotide GTR model).
See Yap, Lindsay, Easteal and Huttley, Mol Biol Evol, In press."""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'CNFGTR',
predicates = _gtr_preds+[_omega],
mprob_model='conditional',
**kw)
def H04GGK(**kw):
"""Huttley 2004 CpG substitution model. Includes a general term for
substitutions to or from CpG's and an adjustment for CpG transitions.
see, GA Huttley. Mol Biol Evol, 21(9):1760-8"""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'H04GGK',
predicates = [_cg, _cg_k, _kappa, _omega],
mprob_model = 'tuple',
**kw)
def MG94GTR(**kw):
"""Muse and Gaut 1994 codon substitution model, GTR variant (with params
analagous to the nucleotide GTR model)
see, Muse and Gaut, 1994, Mol Biol Evol, 11, 715-24"""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'MG94',
predicates = _gtr_preds+[_omega],
mprob_model='monomer',
**kw)
def test_binned_gamma_ordered_param(self):
"""rate is gamma distributed omega follows"""
submod = substitution_model.Codon(predicates={
'kappa': 'transition',
'omega': 'replacement'
},
ordered_param='rate',
partitioned_params='omega',
distribution='gamma',
mprob_model='tuple')
lf = self._makeLikelihoodFunction(submod, self.alignment, bins=3)
values = lf.getParamValueDict(['bin'])['omega_factor'].values()
self.assertEqual(round(sum(values) / len(values), 6), 1.0)
self.assertEqual(len(values), 3)
shape = lf.getParamValue('rate_shape')
def Y98(**kw):
"""Yang's 1998 substitution model, a derivative of the GY94.
see Z Yang. Mol. Biol. Evol., 15(5):568-73, 1998"""
return substitution_model.Codon(
motif_probs = None,
do_scaling = True,
model_gaps = False,
recode_gaps = True,
name = 'Y98',
predicates = {
'kappa' : 'transition',
'omega' : 'replacement',
},
mprob_model = 'tuple',
**kw)
def test_binned_gamma(self):
"""just rate is gamma distributed"""
submod = substitution_model.Codon(
predicates={'kappa': 'transition', 'omega': 'replacement'},
ordered_param='rate', distribution='gamma', mprob_model='tuple')
lf = self._makeLikelihoodFunction(submod, bins=3)
try:
values = list(lf.getParamValueDict(['bin'])['omega_factor'].values())
except KeyError:
# there shouldn't be an omega factor
pass
values = list(lf.getParamValueDict(['bin'])['rate'].values())
obs = round(sum(values) / len(values), 6)
self.assertEqual(obs, 1.0)
self.assertEqual(len(values), 3)
shape = lf.getParamValue('rate_shape')
def test_codon(self):
"""test a three taxa codon model."""
submod = substitution_model.Codon(equal_motif_probs=True,
do_scaling=False,
motif_probs=None,
predicates={
'kappa': 'transition',
'omega': 'replacement'
},
mprob_model='tuple')
self.par_values.update({'omega': 0.5})
likelihood_function = self._makeLikelihoodFunction(
submod, self.alignment)
for par, val in self.par_values.items():
likelihood_function.setpar(par, val)
likelihood_function.setpar("length", self.length)
evolve_lnL = likelihood_function.testfunction()
self.assertFloatEqual(evolve_lnL, -80.67069614541883)
