def readms1peak(bn):
ms1peaks = []
with open('ms1feature_' + bn + '.txt') as ms1peakfile:
for line in ms1peakfile:
lsp = line.rstrip().split()
if len(lsp) == 5:
ms1peaks.append(Peak(*[float(x) for x in lsp], float(lsp[0])))
ms1peaks.sort()
iso_diff = 1.00335
single = 0
doub = 0
peak_double = set()
for ii in range(len(ms1peaks)):
peak0 = ms1peaks[ii]
err_bd = min(.01, bound_ppm(peak0.mz * ms1ppm))
pos0 = bisect_left(ms1peaks, (peak0.mz - iso_diff - err_bd, ))
pos1 = bisect_left(ms1peaks, (peak0.mz - iso_diff + err_bd, ), lo=pos0)
if pos0 != pos1:
peak1 = min(ms1peaks[pos0:pos1],
key=lambda p: abs(peak0.rt - p.rt))
if abs(peak0.rt - peak1.rt) < 1 and peak0.auc < peak1.auc:
ms1peaks[ii] = Peak(*peak0[:-1], peak1.mmz)
if peak1.mmz == peak1.mz:
single += 1
return ms1peaks, peak_double
def print_tab(lib_ent):
with open('quant_rex.txt', 'w') as quant_auc:
quant_auc.write(
'name\tISF\tadduct\tfeature_m/z(library)\tRT(library)\tfeature_m/z(experimental median)\t%detected\t'
)
quant_auc.write('\t'.join(x[:-5] for x in mzML_files) + '\t')
quant_auc.write('\t'.join('RT_' + x[:-5] for x in mzML_files) + '\n')
for ent in lib_ent:
ent_p = []
for nn, mzML_file in enumerate(mzML_files):
ms1peaks = ms1peaks_dict[mzML_file]
bd = bound_ppm(ent.Mmass * ms1ppm)
pos0 = bisect_left(ms1peaks, (ent.Mmass - bd, ))
pos1 = bisect_left(ms1peaks, (ent.Mmass + bd, ))
if ent.rt != 'NA':
peak = [
p for p in ms1peaks[pos0:pos1]
if abs(p.rt - ent.rt) < RT_shift
]
else:
peak = ms1peaks[pos0:pos1]
if peak:
if ent.rt != 'NA':
peak = [max(peak, key=operator.attrgetter('auc'))]
for p in peak:
ent_p.append((nn, p))
line_str = []
if ent_p:
for nn in range(len(mzML_files)):
pos0 = bisect_left([x[0] for x in ent_p], nn)
pos1 = bisect.bisect_right([x[0] for x in ent_p],
nn,
lo=pos0)
line_str.append(( \
','.join(format(dat_n[1].auc,'.1f') for dat_n in ent_p[pos0:pos1]), \
','.join(format(dat_n[1].rt/60,'.2f') for dat_n in ent_p[pos0:pos1]) ))
mzmed = statistics.median(x[1].mz for x in ent_p)
if ent.rt != 'NA':
quant_auc.write(
'{}\t{}\t{}\t{:.5f}\t{:.2f}\t{:.5f}'.format(
ent.name,
'*' if ent.name.startswith('ISF of ') else '',
ent.adduct, ent.Mmass, ent.rt / 60, mzmed))
else:
quant_auc.write('{}\t{}\t{:.5f}\tNA\t{:.5f}'.format(
ent.name, ent.adduct, ent.Mmass, mzmed))
quant_auc.write('\t{:.2f}'.format(
sum((1 if x else 0)
for x, _ in line_str) / len(mzML_files)))
quant_auc.write('\t' + '\t'.join(x for x, _ in line_str) +
'\t' + '\t'.join(x
for _, x in line_str) + '\n')
break
mz_list = [float(x) for x in mz_list]
I_list = [x / max(I_list) * 999. for x in I_list]
all_dat[name_id].append(
(nn, dotp, premz, rt, mz_list, I_list, auc, feat))
lib_dict[name_id] = (dotp, lib_dat)
isf_dict = dict()
for name_id in all_dat.keys():
name_add = name_id.split('\n')
name, adduct = name_add[:-1], name_add[-1]
name0 = name[0]
if name0.startswith('ISF of'):
pmz, rt, mz = [float(x) for x in re.findall("\d+\.\d+", name0)[:3]]
for pmz0, rt0, mz0, adduct0 in isf_dict.keys():
if abs(pmz - pmz0) < bound_ppm(pmz * ms1ppm) and abs(
rt - rt0) < RT_shift and abs(mz - mz0) < bound_ppm(
mz * ms1ppm) and adduct == adduct0:
isf_dict[(pmz0, rt0, mz0, adduct)].append(name_id)
break
else:
isf_dict[(pmz, rt, mz, adduct)] = [name_id]
name_dict0 = dict()
for key, vv in isf_dict.items():
if len(vv) > 1:
pmz_rt_mz = [
tuple(float(x) for x in re.findall("\d+\.\d+", v)[:3]) for v in vv
]
key0 = statistics.median(x for x, _, _ in pmz_rt_mz)
key1 = statistics.median(x for _, x, _ in pmz_rt_mz)
def scoring(cpd):
ent,name=cpd
ms1mz,charge,frag_mz,frag_I,adduct,RT,frag_ann=ent.split(' ',6)
ms1mz=float(ms1mz)
if re.fullmatch("\d+\.\d+",RT): RT=float(RT)
frag_I=[float(x) for x in frag_I.split(',')]
frag_mz=[float(x) for x in frag_mz.split(',')]
frag_ann=[x for x in frag_ann.split(',')]
sorted_I=sorted(zip(frag_I,frag_mz),reverse=True)[:topNfrag]
frag_mz=[x for _,x in sorted_I]
frag_I=[x for x,_ in sorted_I]
if len(frag_I)!=len(frag_mz): print('abb')
adduct=adduct.replace(',','')
name='---'.join(name)
frag_mz_l=[0]*len(frag_mz)
frag_mz_r=[0]*len(frag_mz)
for nn,f_mz in enumerate(frag_mz):
err_bd=.01
frag_mz_l[nn]=f_mz-err_bd
frag_mz_r[nn]=f_mz+err_bd
err_bd=bound_ppm(ms1mz*ms1ppm)
pos0=bisect_left(ms1peaks,(ms1mz-err_bd,))
pos1=bisect_left(ms1peaks,(ms1mz+err_bd,))
pseudo_feat=[]#pseudo feature for entries w/o feature
ms1peaks_match=[x for x in ms1peaks[pos0:pos1] if not isinstance(RT,float) or abs(RT-x.rt)<rt_diff]
if ms2_auc_no_feat and not ms1peaks_match and isinstance(RT,float):#if no feature found and RT in library
pseudo_feat.append(Peak(mz=ms1mz,rt=RT,sc=10,coef=0,auc=0,mmz=ms1mz))
score_peaks=[]
for ms1peak in ms1peaks_match+pseudo_feat:
if not minmz<ms1peak.mz<maxmz:
continue
pos0=bisect_left(startpt,ms1peak.mz)
pos1=bisect_left(end__pt,ms1peak.mz)
if pos0==pos1: # ms1peak in one window only
pos=pos0
else: # take the window whose boundaries are furthest from ms1peak
pos=(pos0 if ms1peak.mz-startpt[pos0-1]>end__pt[pos1]-ms1peak.mz else pos1)
iso=ms2scans[sswath[pos]]
ms2_I=[0]*len(frag_I)
ms2_auc=[0]*len(frag_I)
pfc=[0]*len(frag_I)
rt_l=ms1peak.rt-ms1peak.sc*1.5
rt_r=ms1peak.rt+ms1peak.sc*1.5
ms1rt=rtset[bisect_left(rtset,rt_l):bisect_left(rtset,rt_r)]
p_dict=dict() # highest intensities per scan bounded by m/z
p_area=[x for x in ms1scans[bisect_left(ms1scans,(ms1peak.mz-.01,)):bisect_left(ms1scans,(ms1peak.mz+.01,))] if rt_l<x.rt<rt_r]
for pt in p_area:
if pt.rt not in p_dict or p_dict[pt.rt]<pt.I:
p_dict[pt.rt]=pt.I
p_maxI=[p_dict.get(rt,0.) for rt in ms1rt]
for nn,(f_mz_l,f_mz_r,f_I) in enumerate(zip(frag_mz_l,frag_mz_r,frag_I)):
f_area=[x for x in iso[bisect_left(iso,(f_mz_l,)):bisect_left(iso,(f_mz_r,))] if rt_l<x.rt<rt_r]
if f_area:
f_area_=[x.I for x in f_area if abs(x.rt-ms1peak.rt)<2]
if f_area_:
ms2_I[nn]=max(f_area_)
f_dict=dict() # highest intensities per scan bounded by m/z
for pt in f_area:
if pt.rt not in f_dict or f_dict[pt.rt]<pt.I:
f_dict[pt.rt]=pt.I
ms2_maxI=[f_dict.get(rt,0.) for rt in ms1rt]
pfc[nn]=cos_sim(p_maxI,ms2_maxI)
ms2_auc[nn]=sum((I0+I1)*(rt1-rt0) for rt0,rt1,I0,I1 in zip(ms1rt,ms1rt[1:],ms2_maxI,ms2_maxI[1:]))/2
ms1_auc=sum((I0+I1)*(rt1-rt0) for rt0,rt1,I0,I1 in zip(ms1rt,ms1rt[1:],p_maxI,p_maxI[1:]))/2
ssm=cos_sim0(frag_I,ms2_I)
score_peaks.append((ssm,pfc,ms1peak,ms2_auc,ms1_auc))
if score_peaks:
score_peaks.sort(reverse=True)
max_score_peaks=[score_peaks[0]]
for x in score_peaks[1:]:
if max_score_peaks[0][0]-x[0]<.1:
max_score_peaks.append(x)
else:
break
max_peak=max(max_score_peaks,key=lambda x:x[2].auc) #pick top scoring ms2 for each entry, if score difference is insignificant use auc
return (name,adduct,ms1mz,RT,frag_mz,frag_ann,max_peak)
return False
def mass_matching(jj):
spec, peak = ms2_wp[jj]
adduct_match = []
if "NoMatch" in lib_types[0]:
for adduct0, (mass0, charge0, _) in list(adduct_list.items())[:-1]:
Mmass = ((spec.ms1mz if peak is None else peak.mz) * charge0 -
mass0) / (1 if adduct0[0] == 'M' else int(adduct0[0]))
err_bd = bound_ppm(Mmass * ms1ppm)
pos0 = bisect_left(dlist, (Mmass - err_bd, ))
pos1 = bisect_left(dlist, (Mmass + err_bd, ), lo=pos0)
adduct_match.append((adduct0, (pos0, pos1)))
else:
score_ent = []
premz = (spec.ms1mz if peak is None else peak.mz)
for iiiiii in range(1):
err_bd = bound_ppm(premz * ms1ppm)
pos_0 = bisect_left(lib_ent, (premz - err_bd, ))
pos_1 = bisect_left(lib_ent, (premz + err_bd, ), lo=pos_0)
lib_ent_ = (x for x in lib_ent[pos_0:pos_1]
if x.rt == None or abs(spec.rt - x.rt) < RT_shift)
for ent in lib_ent_:
adduct0, charge0 = ent.adduct, ent.charge
ms2_I = []
ent_I = []
xfrag = set()
for nn, (f_mz, f_I) in enumerate(
(x, y) for x, y in zip(ent.mz, ent.I)
if (charge0 * premz - x) > 3.3):
err_bd = .01
pos0 = bisect_left(spec.mz, f_mz - err_bd)
pos1 = bisect_left(spec.mz, f_mz + err_bd, lo=pos0)
ent_I.append(f_I)
if pos0 != pos1:
ms2_I.append(max(spec.I[pos0:pos1]))
for i in range(pos0, pos1):
xfrag.add(i)
else:
ms2_I.append(0)
if sum(1 for x in ms2_I if
x > 0) >= min_peaks: #min number of matching peak
for nn, (f_mz, f_I) in enumerate(zip(spec.mz, spec.I)):
if nn not in xfrag and (charge0 * premz -
f_mz) > 3.3:
ms2_I.append(f_I)
ent_I.append(0)
cs = cos_sim(
ent_I, ms2_I
) #*sum(x for x,y in zip(ent_I,ms2_I) if y>0)/sum(ent_I)
if cs > MS2_score:
score_ent.append((adduct0, cs, ent))
if score_ent:
max_score_ent = max(
score_ent,
key=operator.itemgetter(1)) #pick top scoring entry
adduct_match.append(max_score_ent)
for cs, ii in isf_sc[jj]: #if isf, attach data
_, peak1 = ms2_wp[ii]
adduct_match.append(
('?', cs,
Ent(
0, 'ISF of (m/z={:.6f}, rt={:.3f}) {:.6f}'.format(
peak1.mz, peak1.rt, peak.mz), [0], [0], None, None,
None)))
return adduct_match, spec, peak
def ms2_spectrum():
ms2_wp = [] #ms2 with ms1 peak info
for nn, spec in enumerate(ms2scans):
err_bd = bound_ppm(spec.ms1mz * ms1ppm)
pos0 = bisect_left(ms1peaks, (spec.ms1mz - err_bd, ))
pos1 = bisect_left(ms1peaks, (spec.ms1mz + err_bd, ), lo=pos0)
s_peak = None
if pos0 != pos1:
peak = min(ms1peaks[pos0:pos1],
key=lambda p: abs(spec.rt - p.rt))
if abs(spec.rt - peak.rt) < peak.sc * 1.5:
s_peak = peak
if s_peak is None or s_peak.mz == s_peak.mmz: #append monoisotopic peak only
ms2_wp.append((spec, s_peak))
peak_ms2 = dict()
ms2_wp_ = []
for s, p in ms2_wp:
if p is None:
ms2_wp_.append((s, p))
elif p not in peak_ms2 or abs(s.rt - p.rt) < abs(peak_ms2[p].rt -
p.rt):
peak_ms2[p] = s
for p, s in peak_ms2.items():
ms2_wp_.append((s, p))
ms2_wp = sorted(ms2_wp_)
isf_sc = [[] for x in range(len(ms2_wp))]
for ii in range(len(ms2_wp) - 1, -1, -1):
ms2, peak = ms2_wp[ii]
if peak: # and peak.mz==peak.mmz:
for jj in range(bisect_left(ms2_wp, ((ms2.ms1mz, ), ))):
i_ms2, i_p = ms2_wp[jj]
if i_p:
err_bd = min(.01, bound_ppm(i_p.mz * ms2ppm))
pos0 = bisect_left(ms2.mz, i_p.mz - err_bd)
pos1 = bisect_left(ms2.mz, i_p.mz + err_bd, lo=pos0)
if i_p and abs(peak.rt - i_p.rt
) < ISF_rt_diff and pos0 != pos1 and max(
ms2.I) * .1 < max(
ms2.I[pos0:pos1]
): #require ISF to be n% of base peak
ms2_I = []
i_ms2_I = []
xfrag = set()
for f_i, f_mz in (
(x, y)
for x, y in zip(i_ms2.I_all[:10], i_ms2.mz_all)
if y < i_ms2.ms1mz + .01):
err_bd = .01
pos0 = bisect_left(ms2.mz, f_mz - err_bd)
pos1 = bisect_left(ms2.mz, f_mz + err_bd, lo=pos0)
if pos0 != pos1:
i_ms2_I.append(f_i)
ms2_I.append(max(ms2.I[pos0:pos1]))
for i in range(pos0, pos1):
xfrag.add(i)
else:
i_ms2_I.append(f_i)
ms2_I.append(0)
if sum(1 for x in ms2_I if x > 0) > 1:
for nn, (f_mz,
f_I) in enumerate(zip(ms2.mz, ms2.I)):
if nn not in xfrag and f_mz < i_ms2.ms1mz + .01:
ms2_I.append(f_I)
i_ms2_I.append(0)
isf_par0 = [(x, y) for x, y in zip(i_ms2_I, ms2_I)
if y > 0]
isf_par1 = [(x, y) for x, y in zip(i_ms2_I, ms2_I)
if x > 0]
cs=max(cos_sim([x for x,y in isf_par0],[y for x,y in isf_par0]), \
cos_sim([x for x,y in isf_par1],[y for x,y in isf_par1]))
if cs > ISF_score:
isf_sc[jj].append((cs, ii))
for jj, ic in enumerate(isf_sc):
for cs, ii in ic:
print_ms2(*ms2_wp[ii])
print_ms2(*ms2_wp[jj])
ins_f.write('{}\n\n'.format(cs))
return ms2_wp, isf_sc
def print_score(mzML_file):
basename0 = os.path.basename(mzML_file)
print(basename0)
ms2scans = read_ms2('ms2spectra_' + basename0 + '.txt')
bg = param_dict.get("background")
if bg is not None:
ms2scans_bg = read_ms2('ms2spectra_' +
os.path.basename(glob.glob(bg)[0]) + '.txt')
for ii, ms2sc in enumerate(ms2scans):
err_bd = bound_ppm(ms2sc.ms1mz * ms1ppm)
pos0 = bisect_left(ms2scans_bg, (ms2sc.ms1mz - err_bd, ))
pos1 = bisect_left(ms2scans_bg, (ms2sc.ms1mz + err_bd, ), lo=pos0)
if pos1 != pos0:
closest_bg = min(ms2scans_bg[pos0:pos1],
key=lambda x: abs(x.rt - ms2sc.rt))
if abs(closest_bg.rt - ms2sc.rt) < 60:
new_mz = []
new_I = []
for m, i in zip(ms2sc.mz, ms2sc.I):
pos0 = bisect_left(closest_bg.mz, m - err_bd)
pos1 = bisect_left(closest_bg.mz, m + err_bd, lo=pos0)
if pos0 == pos1:
new_mz.append(m)
new_I.append(i)
ms2scans[ii] = Spec(*ms2sc[:2], new_mz, new_I, *ms2sc[-2:])
ms1peaks, peak_double = readms1peak(basename0)
ins_f = open('isf_' + basename0 + '.txt', 'w')
def print_ms2(ms2, p):
ins_f.write('{} {}\n'.format(p.mz, p.rt))
ins_f.write(' '.join(str(x) for x in ms2.mz_all) + '\n')
ins_f.write(' '.join(str(x) for x in ms2.I_all) + '\n')
def ms2_spectrum():
ms2_wp = [] #ms2 with ms1 peak info
for nn, spec in enumerate(ms2scans):
err_bd = bound_ppm(spec.ms1mz * ms1ppm)
pos0 = bisect_left(ms1peaks, (spec.ms1mz - err_bd, ))
pos1 = bisect_left(ms1peaks, (spec.ms1mz + err_bd, ), lo=pos0)
s_peak = None
if pos0 != pos1:
peak = min(ms1peaks[pos0:pos1],
key=lambda p: abs(spec.rt - p.rt))
if abs(spec.rt - peak.rt) < peak.sc * 1.5:
s_peak = peak
if s_peak is None or s_peak.mz == s_peak.mmz: #append monoisotopic peak only
ms2_wp.append((spec, s_peak))
peak_ms2 = dict()
ms2_wp_ = []
for s, p in ms2_wp:
if p is None:
ms2_wp_.append((s, p))
elif p not in peak_ms2 or abs(s.rt - p.rt) < abs(peak_ms2[p].rt -
p.rt):
peak_ms2[p] = s
for p, s in peak_ms2.items():
ms2_wp_.append((s, p))
ms2_wp = sorted(ms2_wp_)
isf_sc = [[] for x in range(len(ms2_wp))]
for ii in range(len(ms2_wp) - 1, -1, -1):
ms2, peak = ms2_wp[ii]
if peak: # and peak.mz==peak.mmz:
for jj in range(bisect_left(ms2_wp, ((ms2.ms1mz, ), ))):
i_ms2, i_p = ms2_wp[jj]
if i_p:
err_bd = min(.01, bound_ppm(i_p.mz * ms2ppm))
pos0 = bisect_left(ms2.mz, i_p.mz - err_bd)
pos1 = bisect_left(ms2.mz, i_p.mz + err_bd, lo=pos0)
if i_p and abs(peak.rt - i_p.rt
) < ISF_rt_diff and pos0 != pos1 and max(
ms2.I) * .1 < max(
ms2.I[pos0:pos1]
): #require ISF to be n% of base peak
ms2_I = []
i_ms2_I = []
xfrag = set()
for f_i, f_mz in (
(x, y)
for x, y in zip(i_ms2.I_all[:10], i_ms2.mz_all)
if y < i_ms2.ms1mz + .01):
err_bd = .01
pos0 = bisect_left(ms2.mz, f_mz - err_bd)
pos1 = bisect_left(ms2.mz, f_mz + err_bd, lo=pos0)
if pos0 != pos1:
i_ms2_I.append(f_i)
ms2_I.append(max(ms2.I[pos0:pos1]))
for i in range(pos0, pos1):
xfrag.add(i)
else:
i_ms2_I.append(f_i)
ms2_I.append(0)
if sum(1 for x in ms2_I if x > 0) > 1:
for nn, (f_mz,
f_I) in enumerate(zip(ms2.mz, ms2.I)):
if nn not in xfrag and f_mz < i_ms2.ms1mz + .01:
ms2_I.append(f_I)
i_ms2_I.append(0)
isf_par0 = [(x, y) for x, y in zip(i_ms2_I, ms2_I)
if y > 0]
isf_par1 = [(x, y) for x, y in zip(i_ms2_I, ms2_I)
if x > 0]
cs=max(cos_sim([x for x,y in isf_par0],[y for x,y in isf_par0]), \
cos_sim([x for x,y in isf_par1],[y for x,y in isf_par1]))
if cs > ISF_score:
isf_sc[jj].append((cs, ii))
for jj, ic in enumerate(isf_sc):
for cs, ii in ic:
print_ms2(*ms2_wp[ii])
print_ms2(*ms2_wp[jj])
ins_f.write('{}\n\n'.format(cs))
return ms2_wp, isf_sc
ms2_wp, isf_sc = ms2_spectrum()
def mass_matching(jj):
spec, peak = ms2_wp[jj]
adduct_match = []
if "NoMatch" in lib_types[0]:
for adduct0, (mass0, charge0, _) in list(adduct_list.items())[:-1]:
Mmass = ((spec.ms1mz if peak is None else peak.mz) * charge0 -
mass0) / (1 if adduct0[0] == 'M' else int(adduct0[0]))
err_bd = bound_ppm(Mmass * ms1ppm)
pos0 = bisect_left(dlist, (Mmass - err_bd, ))
pos1 = bisect_left(dlist, (Mmass + err_bd, ), lo=pos0)
adduct_match.append((adduct0, (pos0, pos1)))
else:
score_ent = []
premz = (spec.ms1mz if peak is None else peak.mz)
for iiiiii in range(1):
err_bd = bound_ppm(premz * ms1ppm)
pos_0 = bisect_left(lib_ent, (premz - err_bd, ))
pos_1 = bisect_left(lib_ent, (premz + err_bd, ), lo=pos_0)
lib_ent_ = (x for x in lib_ent[pos_0:pos_1]
if x.rt == None or abs(spec.rt - x.rt) < RT_shift)
for ent in lib_ent_:
adduct0, charge0 = ent.adduct, ent.charge
ms2_I = []
ent_I = []
xfrag = set()
for nn, (f_mz, f_I) in enumerate(
(x, y) for x, y in zip(ent.mz, ent.I)
if (charge0 * premz - x) > 3.3):
err_bd = .01
pos0 = bisect_left(spec.mz, f_mz - err_bd)
pos1 = bisect_left(spec.mz, f_mz + err_bd, lo=pos0)
ent_I.append(f_I)
if pos0 != pos1:
ms2_I.append(max(spec.I[pos0:pos1]))
for i in range(pos0, pos1):
xfrag.add(i)
else:
ms2_I.append(0)
if sum(1 for x in ms2_I if
x > 0) >= min_peaks: #min number of matching peak
for nn, (f_mz, f_I) in enumerate(zip(spec.mz, spec.I)):
if nn not in xfrag and (charge0 * premz -
f_mz) > 3.3:
ms2_I.append(f_I)
ent_I.append(0)
cs = cos_sim(
ent_I, ms2_I
) #*sum(x for x,y in zip(ent_I,ms2_I) if y>0)/sum(ent_I)
if cs > MS2_score:
score_ent.append((adduct0, cs, ent))
if score_ent:
max_score_ent = max(
score_ent,
key=operator.itemgetter(1)) #pick top scoring entry
adduct_match.append(max_score_ent)
for cs, ii in isf_sc[jj]: #if isf, attach data
_, peak1 = ms2_wp[ii]
adduct_match.append(
('?', cs,
Ent(
0, 'ISF of (m/z={:.6f}, rt={:.3f}) {:.6f}'.format(
peak1.mz, peak1.rt, peak.mz), [0], [0], None, None,
None)))
return adduct_match, spec, peak
ms1scans, rtall = read_scans(basename0)
def print_ann(ann_, adduct, spec, peak, name):
ann_.write('NAME:\n')
ann_.write(name + '\n')
adduct_c = adduct_list.get(adduct[0], ('', '', ''))
if adduct[0] == '?': #if isf
ann_.write('ADDUCT: -\n')
else:
ann_.write('ADDUCT: {} {}{}\n'.format(adduct[0], adduct_c[1],
adduct_c[2]))
ann_.write('TARGET_M/Z, FEATURE_M/Z: {:.6f}'.format(spec.ms1mz))
ann_.write(', no_ms1_feature_detected' if peak is None else ', ' +
format(peak.mz, '.6f'))
ann_.write('\n')
ann_.write('SCAN_START_TIME, RT: {:.3f}'.format(spec.rt))
if peak is None:
rt_l, rt_r = spec.rt - 10, spec.rt + 10
p_area = [
x for x in ms1scans[bisect_left(ms1scans, (
spec.ms1mz - .01, )):bisect_left(ms1scans, (spec.ms1mz +
.01, ))]
if rt_l < x.rt < rt_r
]
else:
rt_l, rt_r = peak.rt - peak.sc * 1.5, peak.rt + peak.sc * 1.5
p_area = [
x for x in ms1scans[bisect_left(ms1scans, (
peak.mz - .01, )):bisect_left(ms1scans, (peak.mz + .01, ))]
if rt_l < x.rt < rt_r
]
ms1rt = rtall[bisect_left(rtall, rt_l):bisect_left(rtall, rt_r)]
p_dict = dict() # highest intensities per scan bounded by m/z
for pt in p_area:
if pt.rt not in p_dict or p_dict[pt.rt] < pt.I:
p_dict[pt.rt] = pt.I
p_maxI = [p_dict.get(rt, 0.) for rt in ms1rt]
ms1_auc = sum((I0 + I1) * (rt1 - rt0) for rt0, rt1, I0, I1 in zip(
ms1rt, ms1rt[1:], p_maxI, p_maxI[1:])) / 2
ann_.write(', no_ms1_feature_detected' if peak is None else ', ' +
format(peak.rt, '.3f'))
ann_.write('\n')
if len(adduct) == 3:
dotp = adduct[1]
elif len(adduct) == 2:
dotp = False
ann_.write('PEAK_AREA: ' + str(ms1_auc) + '\n')
ann_.write('DOT_PRODUCT: ' + (format(dotp, '.3f') if dotp else '-') +
'\n')
ann_.write('EXPERIMENTAL_SPECTRUM:\n')
for i, mz in zip(spec.I_all, spec.mz_all):
ann_.write('{:.6f} {:.6g}\n'.format(mz, i))
def rec_all():
uk_count = 1
if "NoMatch" not in lib_types[0]:
una_ = open(
'una_' + '_'.join(lib_types) + '_' + basename0 + '.txt', 'w')
id_quant_ma = collections.defaultdict(list)
for adduct_match, spec, peak in map(mass_matching, range(len(ms2_wp))):
for adduct in adduct_match:
if len(adduct) == 3: # if lipidblast
name0 = adduct[2].name
elif len(adduct) == 2: #if nomatch
pos0, pos1 = adduct[1]
nameset = {x for _, x in dlist[pos0:pos1]}
name0 = '\n'.join(sorted(nameset))
id_quant_ma[name0, adduct[0]].append((adduct, spec, peak))
if not adduct_match and "NoMatch" not in lib_types[0]:
una_.write("NAME: unknown_{} {} MS1 feature\n".format(
uk_count, ('with' if peak else 'no')))
uk_count += 1
una_.write("PRECURSORMZ: {:.6f}\n".format(spec.ms1mz))
if ispos:
una_.write(
"PRECURSORTYPE: [M+H]+ [M+2H]2+ [M+Na]+ [M+NH4]+ [M-H2O+H]+\n"
)
else:
una_.write(
"PRECURSORTYPE: [M-H]- [M-2H]2- [M-H2O-H]- [M+HCOO]- [M+Cl]-\n"
)
una_.write("RETENTIONTIME: {:.3f}\n".format(spec.rt / 60))
thres = bpfilter[0] * max(spec.I_all)
I_mz_list = [(i, mz) for i, mz in zip(spec.I_all, spec.mz_all)
if i > thres]
I_mz_list = I_mz_list[:bpfilter[1]]
una_.write("Num Peaks: {}\n".format(len(I_mz_list)))
for i, mz in I_mz_list:
una_.write('{:.6f} {:.6g}\n'.format(mz, i))
una_.write('\n')
return id_quant_ma
id_quant_ma = rec_all()
def print_all(id_quant):
annotated_ = open(
'ann_' + '_'.join(lib_types) + '_' + basename0 + '.txt', 'w')
for (name, adductid), adducts in id_quant.items():
for adduct, spec, peak in adducts:
print_ann(annotated_, adduct, spec, peak, name)
if len(adduct) == 3:
ent = adduct[2]
annotated_.write('LIBRARY_SPECTRUM:\n')
for mz, i in zip(ent.mz, ent.I):
annotated_.write('{} {}\n'.format(mz, i))
annotated_.write('\n')
print_all(id_quant_ma)