def get_lists_of_mzs(sf):
try:
isotope_ms = pyisocalc.isodist(sf,plot=False,sigma=0.01,charges=-2,resolution=100000.0,do_centroid=False)
mzlist = list(isotope_ms.get_mzs())
intenslist = list(isotope_ms.get_intensities())
mzs_list, intensities_list, indices_list = gradient(isotope_ms.get_mzs(), isotope_ms.get_intensities(), max_output=-1, weighted_bins=0)
indices_list = [i if intenslist[i] > intenslist[i+1] else i+1 for i in indices_list]
mzs_list = [mzlist[i] for i in indices_list]
intensities_list = [intenslist[i] for i in indices_list]
min_i = np.min([ i for i in xrange(len(intenslist)) if intenslist[i] > 0.01])
max_i = np.max([ i for i in xrange(len(intenslist)) if intenslist[i] > 0.01])
return {
"isodist_mzs" : mzlist[min_i:max_i],
"isodist_int" : intenslist[min_i:max_i],
"grad_mzs" : list(mzs_list),
"grad_int" : list(intensities_list),
"grad_ind" : list(indices_list - min_i) }
except:
return {
"isodist_mzs" : [],
"isodist_int" : [],
"grad_mzs" : [],
"grad_int" : [],
"grad_ind" : []
}
def load_queries(self):
config = self.config
db_filename = config['file_inputs']['database_file']
db_dump_folder = config['file_inputs']['database_load_folder']
isocalc_sig = config['isotope_generation']['isocalc_sig']
isocalc_resolution = config['isotope_generation']['isocalc_resolution']
if len(config['isotope_generation']['charge']) > 1:
print 'Warning: only first charge state currently accepted'
charge = int('{}{}'.format(config['isotope_generation']['charge'][0]['polarity'], config['isotope_generation']['charge'][0]['n_charges'])) #currently only supports first charge!!
self.adducts=[a['adduct'] for a in config['isotope_generation']['adducts']]
# Read in molecules
self.sum_formulae = [l.strip() for l in open(db_filename).readlines()]
# Check if already generated and load if possible, otherwise calculate fresh
db_name = os.path.splitext(os.path.basename(db_filename))[0]
self.mz_list={}
for adduct in self.adducts:
for sum_formula in self.sum_formulae:
isotope_ms = pyisocalc.isodist(sum_formula + adduct,
plot=False,
sigma=isocalc_sig,
charges=charge,
resolution=isocalc_resolution,
do_centroid=True)
if not sum_formula in self.mz_list:
self.mz_list[sum_formula] = {}
mzs, ints = map(np.array, isotope_ms.get_spectrum(source='centroids'))
order = ints.argsort()[::-1]
self.mz_list[sum_formula][adduct] = (mzs[order], ints[order])
def make_sf_adduct_optimusfilter(sum_formulae,adducts,output_filename,sigma=0.001,resolution=10000,charge=1):
from pyMS.pyisocalc import pyisocalc
# Extract variables from config dict
# Check if already genrated and load if possible, otherwise calculate fresh
with open(output_filename,'a') as f_out:
for sum_formula in sum_formulae:
#print sum_formula
for adduct in adducts:
try:
sf = pyisocalc.complex_to_simple(sum_formula+adduct)
if sf is None: # not possible to form adduct
continue
isotope_ms = pyisocalc.isodist(sf, plot=False, sigma=sigma, charges=charge,
resolution=resolution)
except KeyError as e:
if str(e).startswith("KeyError:"):
print str(e)
continue
except ValueError as e:
if str(e).startswith("Element not recognised"):
print str(e)
continue
except:
print sf=="", sum_formula, adduct
raise
f_out.write("{} [M{}],-1,{}\n".format(sum_formula,adduct,isotope_ms.get_spectrum(source='centroids')[0][0]))
def calcualte_isotope_patterns(sum_formulae,adducts='',isocalc_sig=0.01,isocalc_resolution = 200000.,isocalc_do_centroid = True, charge='1'):
### Generate a mz list of peak centroids for each sum formula with the given adduct
# todo - parse sum formula and adduct properly so that subtractions (losses) can be utilised (this code already exists somewhere)
mz_list={}
for n,sum_formula in enumerate(sum_formulae):
isotope_ms = pyisocalc.isodist(sum_formula+adduct,plot=False,sigma=isocalc_sig,charges=charge,resolution=isocalc_resolution,do_centroid=isocalc_do_centroid)
if not sum_formula in mz_list:
mz_list[sum_formula] = {}
mz_list[sum_formula][adduct] = isotope_ms.get_spectrum(source='centroids')
return mz_list
def show_images_get():
dataset = bottle.request.params.get('dataset', app.paths.iterkeys().next())
formula = bottle.request.params.get('formula', '')
tolerance = float(bottle.request.params.get('tolerance', 5.0))
resolution = float(bottle.request.params.get('resolution', 1e5))
selected_adduct = bottle.request.params.get('adduct', 'H')
hs_removal = bottle.request.GET.get('hs_removal', False)
k = int(bottle.request.params.get('npeaks', 4))
if hs_removal == 'on':
hs_removal = True
pts = float(bottle.request.params.get('pts', 10))
cutoff = float(bottle.request.params.get('pyisocalc_cutoff', 1e-3))
adducts = ['H', 'K', 'Na']
isotope_patterns = {}
for adduct in adducts:
sf = pyisocalc.SumFormulaParser.parse_string(formula + adduct)
raw_pattern = pyisocalc.isodist(sf, cutoff)
fwhm = raw_pattern.get_spectrum()[0][0] / resolution
pattern = pyisocalc.apply_gaussian(raw_pattern, fwhm, pts, exact=True)
mzs, intensities = map(np.array, pattern.get_spectrum(source='centroids'))
if len(mzs) > k:
order = intensities.argsort()[::-1]
mzs = mzs[order][:k]
intensities = intensities[order][:k]
order = mzs.argsort()
mzs = mzs[order]
intensities = intensities[order]
datacube = app.get_datacube(dataset, mzs, tolerance)
if hs_removal:
for img in datacube.xic:
if len(img) > 0:
pc = np.percentile(img, 99)
img[img > pc] = pc
chaos = measure_of_chaos(datacube.xic_to_image(0), 30, overwrite=False)
iso_corr = isotope_pattern_match(datacube.xic, intensities)
img_corr = 1.0 # return 1 if there's a single peak
if len(intensities[1:]) > 1:
img_corr = isotope_image_correlation(datacube.xic, weights=intensities[1:])
stats = {'measure of chaos': chaos,
'image correlation score': img_corr,
'isotope pattern score': iso_corr}
isotope_patterns[adduct] = (mzs, intensities, stats)
return bottle.template('show_images', hs_removal=hs_removal,
isotope_patterns=isotope_patterns, formula=formula, selected_adduct=selected_adduct,
pretty_formula=re.sub(r"(\d+)", r"<sub>\1</sub>", formula),
resolution=resolution, tol=tolerance, datasets=app.paths.keys(),
npeaks=k, selected_dataset=dataset)
def calculate_isotope_patterns(sum_formulae,
adduct='',
isocalc_sig=0.01,
isocalc_resolution=200000.,
isocalc_do_centroid=True,
charge=1,
verbose=True):
from pyMS.pyisocalc import pyisocalc
### Generate a mz list of peak centroids for each sum formula with the given adduct
mz_list = {}
for n, sum_formula in enumerate(sum_formulae):
try:
if verbose:
print sum_formula, adduct
sf = pyisocalc.complex_to_simple(sum_formula + adduct)
except KeyError as e:
if str(e).startswith("KeyError: "):
print str(e)
continue
except ValueError as e:
if str(e).startswith("Element not recognised"):
print str(e)
continue
else:
print sum_formula, adduct
raise
if sf == None: #negative atoms as a result of simplification
print 'negative adduct for {} : {}'.format(sum_formula, adduct)
continue
try:
isotope_ms = pyisocalc.isodist(sf,
plot=False,
sigma=isocalc_sig,
charges=charge,
resolution=isocalc_resolution,
do_centroid=isocalc_do_centroid)
except KeyError as e:
if str(e).startswith("KeyError: "):
print str(e)
continue
if not sum_formula in mz_list:
mz_list[sum_formula] = {}
mz_list[sum_formula][adduct] = isotope_ms.get_spectrum(
source='centroids')
return mz_list
def calculate_isotope_patterns(sum_formulae,adduct,isocalc_resolution,isocalc_do_centroid = True, charge=1):
### Generate a mz list of peak centroids for each sum formula with the given adduct
mz_list={}
for n, sum_formula in enumerate(sum_formulae):
sf = pyisocalc.SumFormulaParser.parse_string(str(sum_formula + adduct))
raw_pattern = pyisocalc.isodist(sf, cutoff=1e-4, charge=charge)
mz = raw_pattern.get_spectrum()[0][0]
# if mz < 200 or mz > 2000:
# continue
fwhm = mz / isocalc_resolution # TODO: resolution = resolution(mz)
isotope_ms = pyisocalc.apply_gaussian(raw_pattern, fwhm, exact=False)
if not sum_formula in mz_list:
mz_list[sum_formula] = {}
mzs, intensities = isotope_ms.get_spectrum(source='centroids')
order = intensities.argsort()[::-1][:5]
mz_list[sum_formula][adduct] = (mzs[order], intensities[order])
return mz_list
def generate_patterns(formulas_fn, resolution_func, mz_range):
mz_min, mz_max = mz_range
patterns = {}
adducts = ['H', 'K', 'Na']
formulae = [s.strip() for s in open(formulas_fn).readlines()]
for f in formulae:
for a in adducts:
sf = pyisocalc.SumFormulaParser.parse_string(f + a)
raw_pattern = pyisocalc.isodist(sf, cutoff=1e-4, charge=1)
mz = raw_pattern.get_spectrum()[0][0]
if mz < mz_min or mz > mz_max:
continue
fwhm = mz / resolution_func(mz)
mzs, intensities = pyisocalc.apply_gaussian(raw_pattern, fwhm, exact=False).get_spectrum(source="centroids")
mzs = np.array(mzs)
intensities = np.array(intensities)
order = np.argsort(intensities)[::-1]
patterns[(f, a)] = (mzs[order], intensities[order])
return patterns
def get_lists_of_mzs(sf):
try:
isotope_ms = pyisocalc.isodist(sf,
plot=False,
sigma=0.01,
charges=-2,
resolution=100000.0,
do_centroid=False)
mzlist = list(isotope_ms.get_mzs())
intenslist = list(isotope_ms.get_intensities())
mzs_list, intensities_list, indices_list = gradient(
isotope_ms.get_mzs(),
isotope_ms.get_intensities(),
max_output=-1,
weighted_bins=0)
indices_list = [
i if intenslist[i] > intenslist[i + 1] else i + 1
for i in indices_list
]
mzs_list = [mzlist[i] for i in indices_list]
intensities_list = [intenslist[i] for i in indices_list]
min_i = np.min(
[i for i in xrange(len(intenslist)) if intenslist[i] > 0.01])
max_i = np.max(
[i for i in xrange(len(intenslist)) if intenslist[i] > 0.01])
return {
"isodist_mzs": mzlist[min_i:max_i],
"isodist_int": intenslist[min_i:max_i],
"grad_mzs": list(mzs_list),
"grad_int": list(intensities_list),
"grad_ind": list(indices_list - min_i)
}
except:
return {
"isodist_mzs": [],
"isodist_int": [],
"grad_mzs": [],
"grad_int": [],
"grad_ind": []
}
def calculate_isotope_patterns(sum_formulae, adduct='', isocalc_sig=0.01, isocalc_resolution=200000.,
isocalc_do_centroid=True, charge=1,verbose=True):
from pyMS.pyisocalc import pyisocalc
### Generate a mz list of peak centroids for each sum formula with the given adduct
mz_list = {}
for n, sum_formula in enumerate(sum_formulae):
try:
if verbose:
print sum_formula, adduct
sf = pyisocalc.complex_to_simple(sum_formula+adduct)
except KeyError as e:
if str(e).startswith("KeyError: "):
print str(e)
continue
except ValueError as e:
if str(e).startswith("Element not recognised"):
print str(e)
continue
else:
print sum_formula, adduct
raise
if sf == None: #negative atoms as a result of simplification
print 'negative adduct for {} : {}'.format(sum_formula,adduct)
continue
try:
isotope_ms = pyisocalc.isodist(sf, plot=False, sigma=isocalc_sig, charges=charge,
resolution=isocalc_resolution, do_centroid=isocalc_do_centroid)
except KeyError as e:
if str(e).startswith("KeyError: "):
print str(e)
continue
if not sum_formula in mz_list:
mz_list[sum_formula] = {}
mz_list[sum_formula][adduct] = isotope_ms.get_spectrum(source='centroids')
return mz_list
def load_queries(self):
config = self.config
db_filename = config['file_inputs']['database_file']
db_dump_folder = config['file_inputs']['database_load_folder']
isocalc_sig = config['isotope_generation']['isocalc_sig']
isocalc_resolution = config['isotope_generation']['isocalc_resolution']
if len(config['isotope_generation']['charge']) > 1:
print 'Warning: only first charge state currently accepted'
charge = int('{}{}'.format(
config['isotope_generation']['charge'][0]['polarity'],
config['isotope_generation']['charge'][0]
['n_charges'])) #currently only supports first charge!!
self.adducts = [
a['adduct'] for a in config['isotope_generation']['adducts']
]
# Read in molecules
self.sum_formulae = [l.strip() for l in open(db_filename).readlines()]
# Check if already generated and load if possible, otherwise calculate fresh
db_name = os.path.splitext(os.path.basename(db_filename))[0]
self.mz_list = {}
for adduct in self.adducts:
for sum_formula in self.sum_formulae:
isotope_ms = pyisocalc.isodist(sum_formula + adduct,
plot=False,
sigma=isocalc_sig,
charges=charge,
resolution=isocalc_resolution,
do_centroid=True)
if not sum_formula in self.mz_list:
self.mz_list[sum_formula] = {}
mzs, ints = map(np.array,
isotope_ms.get_spectrum(source='centroids'))
order = ints.argsort()[::-1]
self.mz_list[sum_formula][adduct] = (mzs[order], ints[order])
from pyMS.pyisocalc import pyisocalc
sum_formulae = [l.strip() for l in open("formulae.txt")]
adducts = ['H', 'Na', 'K']
patterns = {}
import os
import cPickle
if os.path.isfile("patterns.pkl"):
patterns = cPickle.load(open("patterns.pkl"))
else:
print "generating patterns..."
for n, sum_formula in enumerate(sum_formulae):
for adduct in adducts:
isotope_ms = pyisocalc.isodist(sum_formula + adduct,
plot=False,
sigma=0.01,
charges=1,
resolution=200000,
do_centroid=True)
if not sum_formula in patterns:
patterns[sum_formula] = {}
patterns[sum_formula][adduct] = isotope_ms.get_spectrum(source='centroids')
with open('patterns.pkl', 'w') as f:
cPickle.dump(patterns, f)
formulas = [k + '+' + a for k in patterns for a in patterns[k]]
masses = [patterns[k][a][0] for k in patterns for a in patterns[k]]
all_masses = np.concatenate(masses)
order = all_masses.argsort()
all_masses = all_masses[order]
mol_indices = np.repeat(np.arange(len(masses)), map(len, masses))[order]
mass_diffs = np.diff(all_masses)