def test_multiple_adducts(self):
fs = DatabaseFormulaSampler(HMDB)
ri = UniformRTAndIntensitySampler(min_rt=100, max_rt=101)
cs = ConstantChromatogramSampler()
adduct_prior_dict = {POSITIVE: {'M+H': 100, 'M+Na': 100, 'M+K': 100}}
cm = ChemicalMixtureCreator(fs,
rt_and_intensity_sampler=ri,
chromatogram_sampler=cs,
adduct_prior_dict=adduct_prior_dict,
adduct_proportion_cutoff=0.0)
n_adducts = len(adduct_prior_dict[POSITIVE])
n_chems = 5
dataset = cm.sample(n_chems, 2)
for c in dataset:
c.isotopes = [(c.mass, 1, "Mono")]
# should be 15 peaks or less all the time
# some adducts might not be sampled if the probability is less than 0.2
controller = SimpleMs1Controller()
ms = IndependentMassSpectrometer(POSITIVE, dataset)
env = Environment(ms, controller, 102, 110, progress_bar=True)
set_log_level_warning()
env.run()
for scan in controller.scans[1]:
assert len(scan.mzs) <= n_chems * n_adducts
def simple_dataset():
ri = UniformRTAndIntensitySampler(min_rt=RT_RANGE[0][0],
max_rt=RT_RANGE[0][1])
hf = DatabaseFormulaSampler(HMDB)
cc = ChemicalMixtureCreator(hf,
rt_and_intensity_sampler=ri,
adduct_prior_dict=ADDUCT_DICT_POS_MH)
d = cc.sample(N_CHEMS, 2)
return d
def test_hmdb_creation(self):
ri = UniformRTAndIntensitySampler(min_rt=RT_RANGE[0][0],
max_rt=RT_RANGE[0][1])
hf = DatabaseFormulaSampler(HMDB,
min_mz=MZ_RANGE[0][0],
max_mz=MZ_RANGE[0][1])
cc = ChemicalMixtureCreator(hf, rt_and_intensity_sampler=ri)
d = cc.sample(N_CHEMS, 2)
check_chems(d)
def test_ms2_mgf(self):
hf = DatabaseFormulaSampler(HMDB,
min_mz=MZ_RANGE[0][0],
max_mz=MZ_RANGE[0][1])
ri = UniformRTAndIntensitySampler(min_rt=RT_RANGE[0][0],
max_rt=RT_RANGE[0][1])
cs = MGFMS2Sampler(MGF_FILE)
cc = ChemicalMixtureCreator(hf,
rt_and_intensity_sampler=ri,
ms2_sampler=cs)
d = cc.sample(N_CHEMS, 2)
check_chems(d)
def test_linked_ms1_ms2_creation(self):
# make a database from an mgf
database = mgf_to_database(MGF_FILE, id_field="SPECTRUMID")
hd = DatabaseFormulaSampler(database)
# ExactMatchMS2Sampler needs to be given the same mgf file
# and both need to use the same field in the MGF as the unique ID
mm = ExactMatchMS2Sampler(MGF_FILE, id_field="SPECTRUMID")
cm = ChemicalMixtureCreator(hd, ms2_sampler=mm)
dataset = cm.sample(N_CHEMS, 2)
# check each chemical to see if it has the correct number of peaks
records = load_mgf(MGF_FILE, id_field="SPECTRUMID")
for chem in dataset:
orig_spec = records[chem.database_accession]
assert len(chem.children) > 0
assert len(orig_spec.peaks) == len(chem.children)
def generate_chems(cls, output_dir, no_chems):
hmdb = load_obj(cls.c.HMDBPATH)
df = DatabaseFormulaSampler(hmdb, min_mz=100, max_mz=1000)
cm = ChemicalMixtureCreator(df,
adduct_prior_dict={POSITIVE: {
"M+H": 1
}})
chemicals = cm.sample(no_chems, 1)
min_rt, max_rt = min(chem.rt for chem in chemicals) * 0.9, max(
chem.rt for chem in chemicals) * 1.1
Path(output_dir).mkdir(exist_ok=True)
with open(os.path.join(output_dir, "rts.txt"), 'w') as rts:
rts.write("{},{}".format(min_rt, max_rt))
save_obj(chemicals, os.path.join(output_dir, "chems.pkl"))
return chemicals, os.path.join(output_dir, "chems.pkl"), os.path.join(
output_dir, "rts.txt")
def run_vimms(no_injections, rt_box_size, mz_box_size):
rt_range = [(0, 1440)]
min_rt, max_rt = rt_range[0]
ionisation_mode, isolation_width = POSITIVE, 1
N, rt_tol, mz_tol, min_ms1_intensity = 10, 15, 10, 5000
min_roi_intensity, min_roi_length, min_roi_length_for_fragmentation = \
500, 3, 3
grid = GridEstimator(
LocatorGrid(min_rt, max_rt, rt_box_size, 0, 3000, mz_box_size),
IdentityDrift())
hmdbpath = os.path.join(os.path.abspath(os.getcwd()), "..", "..", "tests",
"fixtures", "hmdb_compounds.p")
hmdb = load_obj(hmdbpath)
df = DatabaseFormulaSampler(hmdb, min_mz=100, max_mz=1000)
cm = ChemicalMixtureCreator(df, adduct_prior_dict={POSITIVE: {"M+H": 1}})
chemicals = cm.sample(2000, 1)
boxes = []
for i in range(no_injections):
mz_noise = GaussianPeakNoise(0.1)
mass_spec = IndependentMassSpectrometer(POSITIVE, chemicals,
mz_noise=mz_noise)
controller = NonOverlapController(
ionisation_mode, isolation_width, mz_tol, min_ms1_intensity,
min_roi_intensity,
min_roi_length, N, grid, rt_tol=rt_tol,
min_roi_length_for_fragmentation=min_roi_length_for_fragmentation
)
env = Environment(mass_spec, controller, min_rt, max_rt,
progress_bar=True)
set_log_level_warning()
env.run()
boxes.append(
[r.to_box(0.01, 0.01) for r in controller.roi_builder.get_rois()])
return boxes
def test_multiple_chems(self):
hf = DatabaseFormulaSampler(HMDB,
min_mz=MZ_RANGE[0][0],
max_mz=MZ_RANGE[0][1])
ri = UniformRTAndIntensitySampler(min_rt=RT_RANGE[0][0],
max_rt=RT_RANGE[0][1])
cc = ChemicalMixtureCreator(hf, rt_and_intensity_sampler=ri)
d = cc.sample(N_CHEMS, 2)
group_list = ['control', 'control', 'case', 'case']
group_dict = {
'case': {
'missing_probability': 0,
'changing_probability': 0
}
}
# missing noise
peak_noise = NoPeakNoise()
mm = MultipleMixtureCreator(d,
group_list,
group_dict,
intensity_noise=peak_noise)
cl = mm.generate_chemical_lists()
for c in cl:
check_chems(c)
# with these settings all chemicals should be in all lists with identical intensities
originals = [f.base_chemical for f in c]
assert len(set(originals)) == len(d)
for f in c:
assert f.max_intensity == f.base_chemical.max_intensity
group_dict = {
'case': {
'missing_probability': 1.,
'changing_probability': 0
}
}
mm = MultipleMixtureCreator(d,
group_list,
group_dict,
intensity_noise=peak_noise)
cl = mm.generate_chemical_lists()
for i, c in enumerate(cl):
if group_list[i] == 'case':
assert len(c) == 0
# test the case that if the missing probability is 1 all are missing
group_dict = {
'case': {
'missing_probability': 1.,
'changing_probability': 0
}
}
mm = MultipleMixtureCreator(d,
group_list,
group_dict,
intensity_noise=peak_noise)
cl = mm.generate_chemical_lists()
for i, c in enumerate(cl):
if group_list[i] == 'case':
assert len(c) == 0
# test the case that changing probablity is 1 changes everything
group_dict = {
'case': {
'missing_probability': 0.,
'changing_probability': 1.
}
}
mm = MultipleMixtureCreator(d,
group_list,
group_dict,
intensity_noise=peak_noise)
cl = mm.generate_chemical_lists()
for i, c in enumerate(cl):
if group_list[i] == 'case':
for f in c:
assert not f.max_intensity == f.base_chemical.max_intensity
parser.add_argument('--output_swath_file',
dest='output_swath_file',
type=str,
default=None)
parser.add_argument('--print_chems',
dest='print_chems',
action='store_true')
args = parser.parse_args()
formula_database = load_obj(args.formula_database_file)
logger.debug("Loaded {} formulas".format(len(formula_database)))
fs = DatabaseFormulaSampler(formula_database,
min_mz=args.min_mz,
max_mz=args.max_mz)
ri = UniformRTAndIntensitySampler(
min_rt=args.min_rt,
max_rt=args.max_rt,
min_log_intensity=np.log(args.min_ms1_sampling_intensity),
max_log_intensity=np.log(args.max_ms1_sampling_intensity))
cs = UniformMS2Sampler()
cm = ChemicalMixtureCreator(fs,
rt_and_intensity_sampler=ri,
ms2_sampler=cs,
adduct_prior_dict=ADDUCT_DICT_POS_MH)
dataset = cm.sample(args.n_chems, args.ms_levels)