pycompound 0.1.0__py3-none-any.whl → 0.1.2__py3-none-any.whl

pycompound/build_library.py

@@ -1,6 +1,4 @@
 
-# this script has a function to extract the mass spectra from an mgf, mzML, or cdf file and write them in the necessary format for use in spectral library matching
-
 import netCDF4 as nc
 import numpy as np
 import pandas as pd
@@ -14,7 +12,7 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
     Converts mgf, mzML, cdf, or msp file to the necessary format for spectral library matching.
 
     --input_path: Path to input file (must be mgf, mzML, cdf, or msp file). Mandatory argument.
-    --output_path: Path to output CSV file. Default: current working directory.
+    --output_path: Path to output TXT file. Default: current working directory.
     --is_reference: Boolean flag indicating whether IDs of spectra should be written to output. Only pass true if building a reference library with known compound IDs. Only applicable to mgf and msp files. Options: \'True\', \'False\'. Optional argument. Default: False.
     '''
 
@@ -23,7 +21,6 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
         sys.exit()
 
     if output_path is None:
-        #print('Warning: no output_path specified, so library is written to {Path.cwd()}/build_library.csv')
         tmp = input_path.split('/')
         tmp = tmp[(len(tmp)-1)]
         basename = tmp.split('.')[0]
@@ -34,7 +31,6 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
         print('Error: is_reference must be either \'True\' or \'False\'.')
         sys.exit()
 
-    # determine whether an mgf or a mzML file was passed to --input_path
     last_three_chars = input_path[(len(input_path)-3):len(input_path)]
     last_four_chars = input_path[(len(input_path)-4):len(input_path)]
     if last_three_chars == 'mgf' or last_three_chars == 'MGF':
@@ -50,7 +46,6 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
         sys.exit()
 
 
-    # obtain a list of spectra from the input file
     spectra = []
     if input_file_type == 'mgf':
         with mgf.read(input_path, index_by_scans = True) as reader:
@@ -62,7 +57,6 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
                 spectra.append(spec)
 
 
-    # extract the relevant information from each spectra (i.e m/z ratios and intensities)
     if input_file_type == 'mgf' or input_file_type == 'mzML':
         ids = []
         mzs = []
@@ -128,8 +122,7 @@ def build_library_from_raw_data(input_path=None, output_path=None, is_reference=
                         continue
 
 
-    # write CSV file of spectra for use in spectral library matching
     df = pd.DataFrame({'id':ids, 'mz_ratio':mzs, 'intensity':ints})
-    df.to_csv(output_path, index=False)
+    df.to_csv(output_path, index=False, sep='\t')
 
 

pycompound/plot_spectra.py

@@ -1,6 +1,4 @@
 
-# this script's functions plot a given query spectrum against a given reference spectrum before and after spectrum preprocessing transformations
-
 from .processing import *
 from .similarity_measures import *
 import pandas as pd
@@ -36,7 +34,6 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     --output_path: path to output PDF file containing the plots of the spectra before and after preprocessing transformations. If no argument is passed, then the plots will be saved to the PDF ./spectrum1_{spectrum_ID1}_spectrum2_{spectrum_ID2}_plot.pdf in the current working directory.
     '''
 
-    # load query and reference libraries
     if query_data is None:
         print('\nError: No argument passed to the mandatory query_data. Please pass the path to the CSV file of the query data.')
         sys.exit()
@@ -68,7 +65,6 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
         unique_reference_ids = [str(tmp) for tmp in unique_reference_ids]
 
 
-    ##### process input parameters and ensure they are in a valid format #####
     if spectrum_ID1 is not None:
         spectrum_ID1 = str(spectrum_ID1)
     else:
@@ -177,8 +173,6 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
             spec_tmp = spectrum_ID1
             spectrum_ID1 = spectrum_ID2
             spectrum_ID2 = spec_tmp
-        print(unique_query_ids)
-        print(spectrum_ID1)
         query_idx = unique_query_ids.index(spectrum_ID1)
         reference_idx = unique_reference_ids.index(spectrum_ID2)
         q_idxs_tmp = np.where(df_query.iloc[:,0].astype(str) == unique_query_ids[query_idx])[0]
@@ -192,7 +186,6 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     q_spec_pre_trans[:,1] = q_spec_pre_trans[:,1].astype(float)
     r_spec_pre_trans[:,1] = r_spec_pre_trans[:,1].astype(float)
 
-    # apply transformation to y-axis if relevant
     if y_axis_transformation == 'normalized':
         q_spec_pre_trans[:,1] = q_spec_pre_trans[:,1] / np.max(q_spec_pre_trans[:,1])
         r_spec_pre_trans[:,1] = r_spec_pre_trans[:,1] / np.max(r_spec_pre_trans[:,1])
@@ -208,10 +201,8 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     else:
         ylab = 'Raw Intensity'
 
-    # create the figure
     fig, axes = plt.subplots(nrows=2, ncols=1)
 
-    # plot the untransformed spectra
     plt.subplot(2,1,1)
     plt.vlines(x=q_spec_pre_trans[:,0], ymin=[0]*q_spec_pre_trans.shape[0], ymax=q_spec_pre_trans[:,1], linewidth=3, color='blue', label=f'Spectrum ID 1: {spectrum_ID1}')
     plt.vlines(x=r_spec_pre_trans[:,0], ymin=[0]*r_spec_pre_trans.shape[0], ymax=-r_spec_pre_trans[:,1], linewidth=3, color='red', label=f'Spectrum ID 2: {spectrum_ID2}')
@@ -221,7 +212,6 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     plt.yticks(fontsize=7)
     plt.title('Untransformed Spectra', fontsize=10)
 
-    # get the ranges of m/z and intensity values to display at the bottom of the two plots
     mz_min_tmp_q = round(q_spec[:,0].min(),1)
     mz_min_tmp_r = round(r_spec[:,0].min(),1)
     int_min_tmp_q = round(q_spec[:,1].min(),1)
@@ -235,51 +225,45 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     int_min_tmp = min([int_min_tmp_q,int_min_tmp_r])
     int_max_tmp = max([int_max_tmp_q,int_max_tmp_r])
 
-    # perform the spectrum preprocessing transformations in the order specified
     is_matched = False
     for transformation in spectrum_preprocessing_order:
-        if transformation == 'C' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # centroiding
+        if transformation == 'C' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             q_spec = centroid_spectrum(q_spec, window_size=window_size_centroiding) 
             r_spec = centroid_spectrum(r_spec, window_size=window_size_centroiding) 
-        if transformation == 'M' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # matching
+        if transformation == 'M' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             m_spec = match_peaks_in_spectra(spec_a=q_spec, spec_b=r_spec, window_size=window_size_matching)
             q_spec = m_spec[:,0:2]
             r_spec = m_spec[:,[0,2]]
             is_matched = True
-        if transformation == 'W' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # weight factor transformation
+        if transformation == 'W' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             q_spec[:,1] = wf_transform(q_spec[:,0], q_spec[:,1], wf_mz, wf_intensity)
             r_spec[:,1] = wf_transform(r_spec[:,0], r_spec[:,1], wf_mz, wf_intensity)
-        if transformation == 'L' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # low-entropy transformation
+        if transformation == 'L' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             q_spec[:,1] = LE_transform(q_spec[:,1], LET_threshold, normalization_method=normalization_method)
             r_spec[:,1] = LE_transform(r_spec[:,1], LET_threshold, normalization_method=normalization_method)
-        if transformation == 'N' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # noise removal
+        if transformation == 'N' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             q_spec = remove_noise(q_spec, nr = noise_threshold)
             r_spec = remove_noise(r_spec, nr = noise_threshold)
-        if transformation == 'F' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1: # filtering
+        if transformation == 'F' and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
             q_spec = filter_spec_lcms(q_spec, mz_min = mz_min, mz_max = mz_max, int_min = int_min, int_max = int_max, is_matched = is_matched)
             r_spec = filter_spec_lcms(r_spec, mz_min = mz_min, mz_max = mz_max, int_min = int_min, int_max = int_max, is_matched = is_matched)
 
-    # intensities of query and reference library
     q_ints = q_spec[:,1]
     r_ints = r_spec[:,1]
 
-    # if there is at least one non-zero intensity ion fragment in either spectra, compute their similarity
     if np.sum(q_ints) != 0 and np.sum(r_ints) != 0 and q_spec.shape[0] > 1 and r_spec.shape[1] > 1:
         similarity_score = get_similarity(similarity_measure, q_ints, r_ints, weights, entropy_dimension)
     else:
         similarity_score = 0
 
-    # plot the transformed spectra
     plt.subplot(2,1,2)
 
-    # display warning message if either spectra are empty or have no non-zero intensity ion fragments
     if q_spec.shape[0] > 1:
         if np.max(q_spec[:,1]) == 0 or np.max(r_spec[:,1]) == 0:
             plt.text(0.5, 0.5, 'The query and/or reference spectrum has no non-zero intensities after transformations.\n Change transformation parameters.', ha='center', va='center', fontsize=7, color='black')
             plt.xticks([])
             plt.yticks([])
         else:
-            # apply transformation to y-axis if relevant
             if y_axis_transformation == 'normalized':
                 q_spec[:,1] = q_spec[:,1] / np.max(q_spec[:,1])
                 r_spec[:,1] = r_spec[:,1] / np.max(r_spec[:,1])
@@ -311,7 +295,7 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     plt.figlegend(loc = 'upper center')
     fig.text(0.05, 0.18, f'Similarity Measure: {similarity_measure.capitalize()}', fontsize=7)
     fig.text(0.05, 0.15, f'Similarity Score: {round(similarity_score,4)}', fontsize=7)
-    fig.text(0.05, 0.12, f'Spectrum Preprocessing Order: {''.join(spectrum_preprocessing_order)}', fontsize=7)
+    fig.text(0.05, 0.12, f"Spectrum Preprocessing Order: {''.join(spectrum_preprocessing_order)}", fontsize=7)
     fig.text(0.05, 0.09, f'High Quality Reference Library: {high_quality_reference_library}', fontsize=7)
     fig.text(0.05, 0.06, f'Window Size (Centroiding): {window_size_centroiding}', fontsize=7)
     fig.text(0.05, 0.03, f'Window Size (Matching): {window_size_matching}', fontsize=7)
@@ -320,6 +304,9 @@ def generate_plots_on_HRMS_data(query_data=None, reference_data=None, spectrum_I
     fig.text(0.45, 0.12, f'Noise Threshold: {noise_threshold}', fontsize=7)
     fig.text(0.45, 0.09, f'Weight Factors (m/z,intensity): ({wf_mz},{wf_intensity})', fontsize=7)
     fig.text(0.45, 0.06, f'Low-Entropy Threshold: {LET_threshold}', fontsize=7)
+    if similarity_measure == 'mixture':
+        fig.text(0.45, 0.03, f'Weights for mixture similarity: {weights}', fontsize=7)
+
     plt.savefig(output_path, format='pdf')
 
     if return_plot == True:
@@ -351,7 +338,6 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
     --output_path: path to output PDF file containing the plots of the spectra before and after preprocessing transformations. If no argument is passed, then the plots will be saved to the PDF ./spectrum1_{spectrum_ID1}_spectrum2_{spectrum_ID2}_plot.pdf in the current working directory.
     '''
 
-    # load query and reference libraries
     if query_data is None:
         print('\nError: No argument passed to the mandatory query_data. Please pass the path to the CSV file of the query data.')
         sys.exit()
@@ -381,7 +367,6 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
             unique_reference_ids = df_reference.iloc[:,0].unique()
 
 
-    ##### process input parameters and ensure they are in a valid format #####
     if spectrum_ID1 is not None:
         spectrum_ID1 = str(spectrum_ID1)
     else:
@@ -456,12 +441,10 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
         print(f'Warning: plots will be saved to the PDF ./spectrum1_{spectrum_ID1}_spectrum2_{spectrum_ID2}_plot.pdf in the current working directory.')
         output_path = f'{Path.cwd()}/spectrum1_{spectrum_ID1}_spectrum2_{spectrum_ID2}.pdf'
 
-    # get m/z values
     min_mz = np.min([np.min(df_query.iloc[:,1]), np.min(df_reference.iloc[:,1])])
     max_mz = np.max([np.max(df_query.iloc[:,1]), np.max(df_reference.iloc[:,1])])
     mzs = np.linspace(min_mz,max_mz,(max_mz-min_mz+1))
 
-    # get unique query/reference library IDs; each query/reference ID corresponds to exactly one query/reference mass spectrum
     unique_query_ids = df_query.iloc[:,0].unique().tolist()
     unique_reference_ids = df_reference.iloc[:,0].unique().tolist()
     unique_query_ids = [str(ID) for ID in unique_query_ids]
@@ -493,7 +476,6 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
     q_spec = convert_spec(q_spec,mzs)
     r_spec = convert_spec(r_spec,mzs)
 
-    # get the ranges of m/z and intensity values to display at the bottom of the two plots
     int_min_tmp_q = min(q_spec[q_spec[:,1].nonzero(),1][0])
     int_min_tmp_r = min(r_spec[r_spec[:,1].nonzero(),1][0])
     int_max_tmp_q = max(q_spec[q_spec[:,1].nonzero(),1][0])
@@ -501,13 +483,10 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
     int_min_tmp = int(min([int_min_tmp_q,int_min_tmp_r]))
     int_max_tmp = int(max([int_max_tmp_q,int_max_tmp_r]))
     
-    # create the figure
     fig, axes = plt.subplots(nrows=2, ncols=1)
 
-    # plot the untransformed spectra
     plt.subplot(2,1,1)
 
-    # display warning message if either spectra have no non-zero ion fragments
     if np.max(q_spec[:,1]) == 0 or np.max(r_spec[:,1]) == 0:
         plt.text(0.5, 0.5, 'The query and/or reference spectrum has no non-zero intensities after transformations.\n Change transformation parameters.', ha='center', va='center', fontsize=7, color='black')
         plt.xticks([])
@@ -518,7 +497,6 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
         q_spec_pre_trans[:,1] = q_spec_pre_trans[:,1].astype(float)
         r_spec_pre_trans[:,1] = r_spec_pre_trans[:,1].astype(float)
 
-        # apply transformation to y-axis if relevant
         if y_axis_transformation == 'normalized':
             q_spec_pre_trans[:,1] = q_spec_pre_trans[:,1] / np.max(q_spec_pre_trans[:,1])
             r_spec_pre_trans[:,1] = r_spec_pre_trans[:,1] / np.max(r_spec_pre_trans[:,1])
@@ -542,32 +520,29 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
         plt.title('Untransformed Query and Reference Spectra', fontsize=10)
 
     for transformation in spectrum_preprocessing_order:
-        if transformation == 'W': # weight factor transformation
+        if transformation == 'W':
             q_spec[:,1] = wf_transform(q_spec[:,0], q_spec[:,1], wf_mz, wf_intensity)
             r_spec[:,1] = wf_transform(r_spec[:,0], r_spec[:,1], wf_mz, wf_intensity)
-        if transformation == 'L': # low-entropy transformation
+        if transformation == 'L':
             q_spec[:,1] = LE_transform(q_spec[:,1], LET_threshold, normalization_method)
             r_spec[:,1] = LE_transform(r_spec[:,1], LET_threshold, normalization_method)
-        if transformation == 'N': # noise removal
+        if transformation == 'N':
             q_spec = remove_noise(q_spec, nr = noise_threshold)
             if high_quality_reference_library == False:
                 r_spec = remove_noise(r_spec, nr = noise_threshold)
-        if transformation == 'F': # filtering with respect to mz and/or intensity
+        if transformation == 'F':
             q_spec = filter_spec_gcms(q_spec, mz_min = mz_min, mz_max = mz_max, int_min = int_min, int_max = int_max)
             if high_quality_reference_library == False:
                 r_spec = filter_spec_gcms(r_spec, mz_min = mz_min, mz_max = mz_max, int_min = int_min, int_max = int_max)
 
-    # compute similarity score; if the spectra contain at most one point, their similarity is considered to be 0
     if q_spec.shape[0] > 1:
         similarity_score = get_similarity(similarity_measure, q_spec[:,1], r_spec[:,1], weights, entropy_dimension)
     else:
         similarity_score = 0
 
  
-    # plot the transformed spectra
     plt.subplot(2,1,2)
 
-    # display warning message if either spectra are empty or have no non-zero intensity ion fragments
     if q_spec.shape[0] == 0 or r_spec.shape[0] == 0:
         plt.text(0.5, 0.5, 'The query and/or reference spectrum has no ion fragments left after transformations.\n Change transformation parameters.', ha='center', va='center', fontsize=7, color='black')
         plt.xticks([])
@@ -577,7 +552,6 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
         plt.xticks([])
         plt.yticks([])
     else:
-        # apply transformation to y-axis if relevant
         if y_axis_transformation == 'normalized':
             q_spec[:,1] = q_spec[:,1] / np.max(q_spec[:,1])
             r_spec[:,1] = r_spec[:,1] / np.max(r_spec[:,1])
@@ -601,18 +575,19 @@ def generate_plots_on_NRMS_data(query_data=None, reference_data=None, spectrum_I
         plt.title(f'Transformed Query and Reference Spectra', fontsize=10)
 
 
-    #plt.subplots_adjust(top = 0.8, hspace = 0.7)
     plt.subplots_adjust(top=0.8, hspace=0.92, bottom=0.3)
     plt.figlegend(loc = 'upper center')
     fig.text(0.05, 0.15, f'Similarity Measure: {similarity_measure.capitalize()}', fontsize=7)
     fig.text(0.05, 0.12, f'Similarity Score: {round(similarity_score,4)}', fontsize=7)
-    fig.text(0.05, 0.09, f'Spectrum Preprocessing Order: {''.join(spectrum_preprocessing_order)}', fontsize=7)
+    fig.text(0.05, 0.09, f"Spectrum Preprocessing Order: {''.join(spectrum_preprocessing_order)}", fontsize=7)
     fig.text(0.05, 0.06, f'High Quality Reference Library: {high_quality_reference_library}', fontsize=7)
     fig.text(0.05, 0.03, f'Raw-Scale M/Z Range: [{min_mz},{max_mz}]', fontsize=7)
     fig.text(0.45, 0.15, f'Raw-Scale Intensity Range: [{int_min_tmp},{int_max_tmp}]', fontsize=7)
     fig.text(0.45, 0.12, f'Noise Threshold: {noise_threshold}', fontsize=7)
     fig.text(0.45, 0.09, f'Weight Factors (m/z,intensity): ({wf_mz},{wf_intensity})', fontsize=7)
     fig.text(0.45, 0.06, f'Low-Entropy Threshold: {LET_threshold}', fontsize=7)
+    if similarity_measure=='mixture':
+        fig.text(0.45, 0.03, f'Weights for mixture similarity: {weights}', fontsize=7)
     plt.savefig(output_path, format='pdf')
 
     if return_plot == True:

pycompound/processing.py

@@ -1,6 +1,4 @@
 
-# This script contains the functions used to transform spectra prior to computing similarity scores
-
 from pycompound.build_library import build_library_from_raw_data
 import scipy.stats
 import numpy as np
@@ -165,7 +163,6 @@ def centroid_spectrum(spec, window_size):
 
     spec = spec[np.argsort(spec[:,0])]
 
-    #Fast check is the spectrum needs centroiding
     mz_array = spec[:, 0]
     need_centroid = 0
     if mz_array.shape[0] > 1:
@@ -180,7 +177,6 @@ def centroid_spectrum(spec, window_size):
             mz_delta_allowed = window_size
 
             if spec[i, 1] > 0:
-                #Find left bound for current peak
                 i_left = i - 1
                 while i_left >= 0:
                     mz_delta_left = spec[i, 0] - spec[i_left, 0]
@@ -190,7 +186,6 @@ def centroid_spectrum(spec, window_size):
                         break
                 i_left += 1
 
-                #Find right bound for current peak
                 i_right = i + 1
                 while i_right < spec.shape[0]:
                     mz_delta_right = spec[i_right, 0] - spec[i, 0]
@@ -199,7 +194,6 @@ def centroid_spectrum(spec, window_size):
                     else:
                         break
 
-                #Merge those peaks
                 intensity_sum = np.sum(spec[i_left:i_right, 1])
                 intensity_weighted_sum = np.sum(spec[i_left:i_right, 0] * spec[i_left:i_right, 1])
 
@@ -246,16 +240,13 @@ def match_peaks_in_spectra(spec_a, spec_b, window_size):
         mass_delta = spec_a[a, 0] - spec_b[b, 0]
         
         if mass_delta < -window_size:
-            # Peak only existed in spec a.
             spec_merged.append([spec_a[a, 0], spec_a[a, 1], peak_b_int])
             peak_b_int = 0.
             a += 1
         elif mass_delta > window_size:
-            # Peak only existed in spec b.
             spec_merged.append([spec_b[b, 0], 0., spec_b[b, 1]])
             b += 1
         else:
-            # Peak existed in both spec.
             peak_b_int += spec_b[b, 1]
             b += 1
 

pycompound/similarity_measures.py

@@ -10,7 +10,6 @@ import sys
 
 
 def S_cos(ints_a, ints_b):
-    # Cosine Similarity Measure
     if np.sum(ints_a) == 0 or np.sum(ints_b) == 0:
         return(0)
     else:
@@ -18,12 +17,10 @@ def S_cos(ints_a, ints_b):
 
 
 def ent_renyi(ints, q):
-    # Computes the Renyi entropy of a probability distribution for a given positive entropy dimension q
     return np.log(sum(np.power(ints,q))) / (1-q)
 
 
 def ent_tsallis(ints, q):
-    # Computes the Tsallis entropy of a probability distribution for a given positive entropy dimension q
     return (sum(np.power(ints,q))-1) / (1-q)