예제 #1
0
    def plot_pathway(self,
                     enriched_genes,
                     pathway_id='hsa05322',
                     figurename=None):

        # config figure name
        if not figurename:
            figurename = '%s.pdf' % pathway_id
        assert (figurename.endswith('.pdf'))

        # fetch pathway
        pathway = KGML_parser.read(kegg_get(pathway_id, "kgml"))

        # change color for pathway elements
        for entry in pathway.entries.values():
            possible_gene_names = entry.graphics[0].name
            matched_name = gene_is_enriched(enriched_genes,
                                            possible_gene_names)
            if matched_name:
                entry.graphics[
                    0].bgcolor = self.enriched_box_color  #set box color
                entry.graphics[
                    0].fgcolor = self.enriched_text_color  # set text color
                entry.graphics[0].name = matched_name
            else:
                entry.graphics[0].bgcolor = self.non_enriched_box_color
                entry.graphics[0].fgcolor = self.non_enriched_text_color
                entry.graphics[0].name = entry.graphics[0].name.split(',')[0]

        canvas = KGMLCanvas(pathway,
                            import_imagemap=True,
                            fontsize=self.fontsize)
        canvas.draw(figurename)
        print('Drawn: ', figurename)
        return pathway
def draw_kegg_map(map_id):
    """ Render a local PDF of a KEGG map with the passed map ID
    """
    # Get the background image first
    pathway = KGML_parser.read(kegg_get(map_id, "kgml"))
    canvas = KGMLCanvas(pathway, import_imagemap=True)
    img_filename = "%s.pdf" % map_id
    canvas.draw(img_filename)
def colorCompounds(pathname, cpdlist, size=20):
    pathway = KGML_parser.read(kegg_get(pathname, "kgml"))
    for element in pathway.compounds:
        for graphic in element.graphics:
            if graphic.name in cpdlist:
                graphic.bgcolor = '#ff0000'
                graphic.width = size
                graphic.height = size
    canvas = KGMLCanvas(pathway, import_imagemap=True)
    canvas.draw("%s.pdf" % pathname)
예제 #4
0
 def test_render_KGML_transparency(self):
     """Rendering of KGML to PDF, with color alpha channel."""
     # We test rendering of the original KGML for KO01100,
     # modifying alpha channel for the lipid pathway
     p = self.data
     with open(p[0].infilename) as f:
         pathway = read(f)
         mod_rs = [e for e in pathway.orthologs if
                   len(set(e.name.split()).intersection(self.ko_ids))]
         for r in mod_rs:
             for g in r.graphics:
                 # Modify hex colour directly by appending alpha channel
                 # to hex string
                 g.fgcolor = g.fgcolor + "77"
                 g.width = 20
         kgml_map = KGMLCanvas(pathway)
         kgml_map.draw(p[0].output_stem + "_transparency.pdf")
     # We test rendering of the original KGML for KO3070,
     # modifying the alpha channel for each ortholog entry
     with open(p[1].infilename) as f:
         pathway = read(f)
         orthologs = list(pathway.orthologs)
         # Use Biopython's ColorSpiral to generate colours
         cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5,
                          jitter=0.03)
         colors = cs.get_colors(len(orthologs))
         for o, c in zip(orthologs, colors):
             # Modify color tuples to add alpha channel
             c = c + (0.5, )
             for g in o.graphics:
                 g.bgcolor = c
         kgml_map = KGMLCanvas(pathway)
         pathway.image = p[1].pathway_image
         kgml_map.import_imagemap = p[1].show_pathway_image
         kgml_map.draw(p[1].output_stem + "_transparency.pdf")
예제 #5
0
 def test_render_KGML_modify(self):
     """Rendering of KGML to PDF, with modification."""
     # We test rendering of the original KGML for KO01100,
     # modifying line width for the lipid pathway
     p = self.data
     with open(p[0].infilename) as f:
         pathway = read(f)
         mod_rs = [
             e for e in pathway.orthologs
             if len(set(e.name.split()).intersection(self.ko_ids))
         ]
         for r in mod_rs:
             for g in r.graphics:
                 g.width = 10
         kgml_map = KGMLCanvas(pathway)
         kgml_map.draw(p[0].output_stem + '_widths.pdf')
     # We test rendering of the original KGML for KO3070,
     # modifying the reaction colours for each ortholog entry
     with open(p[1].infilename) as f:
         pathway = read(f)
         orthologs = [e for e in pathway.orthologs]
         # Use Biopython's ColorSpiral to generate colours
         cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5, jitter=0.03)
         colors = cs.get_colors(len(orthologs))
         for o, c in zip(orthologs, colors):
             for g in o.graphics:
                 g.bgcolor = c
         kgml_map = KGMLCanvas(pathway)
         pathway.image = p[1].pathway_image
         kgml_map.import_imagemap = p[1].show_pathway_image
         kgml_map.draw(p[1].output_stem + '_colors.pdf')
예제 #6
0
 def test_render_KGML_basic(self):
     """Basic rendering of KGML: write to PDF without modification."""
     # We test rendering of the original KEGG KGML using only local
     # files.
     for p in self.data:
         with open(p.infilename, 'rU') as f:
             pathway = read(f)
             pathway.image = p.pathway_image
             kgml_map = KGMLCanvas(pathway)
             kgml_map.import_imagemap = p.show_pathway_image
             kgml_map.draw(p.output_stem + '_original.pdf')
예제 #7
0
 def test_render_KGML_basic(self):
     """Basic rendering of KGML: write to PDF without modification."""
     # We test rendering of the original KEGG KGML using only local
     # files.
     for p in self.data:
         with open(p.infilename, 'rU') as f:
             pathway = read(f)
             pathway.image = p.pathway_image
             kgml_map = KGMLCanvas(pathway)
             kgml_map.import_imagemap = p.show_pathway_image
             kgml_map.draw(p.output_stem + '_original.pdf')
    def test_render_KGML_import_map(self):
        """Basic rendering of KGML: use imported imagemap.

        Uses the URL indicated in the .xml file.

        This test may fail if the imagemap is not available (e.g. if
        there is not a web connection), and may look odd if the remote
        imagemap has changed since the local KGML file was downloaded.
        """
        # We test rendering of the original KEGG KGML using imported files
        for p in self.data:
            with open(p.infilename) as f:
                pathway = read(f)
                kgml_map = KGMLCanvas(pathway, import_imagemap=True)
                kgml_map.draw(p.output_stem + "_importmap.pdf")
    def test_render_KGML_import_map(self):
        """Basic rendering of KGML: use imported imagemap

        Uses the URL indicated in the .xml file.

        This test may fail if the imagemap is not available (e.g. if
        there is not a web connection), and may look odd if the remote
        imagemap has changed since the local KGML file was downloaded.
        """
        # We test rendering of the original KEGG KGML using imported files
        for p in self.data:
            with open(p.infilename, 'rU') as f:
                pathway = read(f)
                kgml_map = KGMLCanvas(pathway, import_imagemap=True)
                kgml_map.draw(p.output_stem + '_importmap.pdf')
예제 #10
0
 def test_render_KGML_modify(self):
     """ Rendering of KGML to PDF, with modification.
     """
     # We test rendering of the original KGML for KO01100,
     # modifying line width for the lipid pathway
     p = self.data
     with open(p[0].infilename) as f:
         pathway = read(f)
         mod_rs = [e for e in pathway.orthologs if
                 len(set(e.name.split()).intersection(self.ko_ids))]
         for r in mod_rs:
             for g in r.graphics:
                 g.width = 10
         kgml_map = KGMLCanvas(pathway)
         kgml_map.draw(p[0].output_stem + '_widths.pdf')
     # We test rendering of the original KGML for KO3070,
     # modifying the reaction colours for each ortholog entry
     with open(p[1].infilename) as f:
         pathway = read(f)
         orthologs = [e for e in pathway.orthologs]
         # Use Biopython's ColorSpiral to generate colours
         cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5, 
                          jitter=0.03)
         colors = cs.get_colors(len(orthologs))
         for o, c in zip(orthologs, colors):
             for g in o.graphics:
                 g.bgcolor = c
         kgml_map = KGMLCanvas(pathway)
         pathway.image = p[1].pathway_image
         kgml_map.import_imagemap = p[1].show_pathway_image
         kgml_map.draw(p[1].output_stem + '_colors.pdf')
예제 #11
0
 def test_render_KGML_transparency(self):
     """Rendering of KGML to PDF, with color alpha channel."""
     # We test rendering of the original KGML for KO01100,
     # modifying alpha channel for the lipid pathway
     p = self.data
     with open(p[0].infilename) as f:
         pathway = read(f)
         mod_rs = [e for e in pathway.orthologs if
                   len(set(e.name.split()).intersection(self.ko_ids))]
         for r in mod_rs:
             for g in r.graphics:
                 # Modify hex colour directly by appending alpha channel
                 # to hex string
                 g.fgcolor = g.fgcolor + "77"
                 g.width = 20
         kgml_map = KGMLCanvas(pathway)
         kgml_map.draw(p[0].output_stem + '_transparency.pdf')
     # We test rendering of the original KGML for KO3070,
     # modifying the alpha channel for each ortholog entry
     with open(p[1].infilename) as f:
         pathway = read(f)
         orthologs = [e for e in pathway.orthologs]
         # Use Biopython's ColorSpiral to generate colours
         cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5,
                          jitter=0.03)
         colors = cs.get_colors(len(orthologs))
         for o, c in zip(orthologs, colors):
             # Modify color tuples to add alpha channel
             c = c + (0.5, )
             for g in o.graphics:
                 g.bgcolor = c
         kgml_map = KGMLCanvas(pathway)
         pathway.image = p[1].pathway_image
         kgml_map.import_imagemap = p[1].show_pathway_image
         kgml_map.draw(p[1].output_stem + '_transparency.pdf')
예제 #12
0
 def getPic(self, pidpath, pathwaydir):
     """
     输入文件pid.txt,输出文件夹pathways,作图
     """
     fs = gridfs.GridFS(self.mongodb)
     f = open(pidpath)
     if not os.path.exists(pathwaydir):
         os.makedirs(pathwaydir)
     for i in f:
         if i:
             i = i.strip('\n').split('\t')
             pid = i[0]
             koid = i[1].split(';')
             l = []
             kgml_path = os.path.join(os.getcwd(), "pathway.kgml")
             png_path = os.path.join(os.getcwd(), "pathway.png")
             if os.path.exists(kgml_path) and os.path.exists(png_path):
                 os.remove(kgml_path)
                 os.remove(png_path)
             with open("pathway.kgml",
                       "w+") as k, open("pathway.png", "w+") as p:
                 result = self.png_coll.find_one({"pathway_id": pid})
                 if result:
                     kgml_id = result['pathway_ko_kgml']
                     png_id = result['pathway_ko_png']
                     k.write(fs.get(kgml_id).read())
                     p.write(fs.get(png_id).read())
             p_kgml = KGML_parser.read(open("pathway.kgml"))
             p_kgml.image = png_path
             for ko in koid:
                 for degree in p_kgml.entries.values():
                     if re.search(ko, degree.name):
                         l.append(degree.id)
                 for n in l:
                     for graphic in p_kgml.entries[n].graphics:
                         graphic.fgcolor = '#CC0000'
                 canvas = KGMLCanvas(p_kgml, import_imagemap=True)
                 canvas.draw(pathwaydir + '/' + pid + '.pdf')
     print "getPic finished!!!"
예제 #13
0
def map2highlighted_map(map_id,
                        ko_list,
                        ko2freq,
                        biodb,
                        outpath='test.pdf',
                        taxon_id=False,
                        n_species=60):
    import re
    from chlamdb.biosqldb import shell_command
    from Bio.Graphics.KGML_vis import KGMLCanvas
    from Bio.Graphics import KGML_vis
    import urllib.request
    from Bio.KEGG.KGML.KGML_pathway import Pathway, Reaction, Relation
    import Bio.KEGG.KGML.KGML_pathway
    from Bio.KEGG.KGML import KGML_parser
    from Bio.Graphics.ColorSpiral import ColorSpiral
    import matplotlib.cm as cm
    from matplotlib.colors import rgb2hex
    import matplotlib as mpl

    values = [float(i) for i in ko2freq.values()]

    norm = mpl.colors.Normalize(vmin=0, vmax=n_species)
    cmap = cm.OrRd
    cmap2 = cm.Greens
    m = cm.ScalarMappable(norm=norm, cmap=cmap)
    m2 = cm.ScalarMappable(norm=norm, cmap=cmap2)

    url_template = 'http://rest.kegg.jp/get/%s/kgml' % re.sub(
        'map', 'ko', map_id)
    print(url_template)
    f = urllib.request.urlopen(url_template)
    from Bio.Graphics import KGML_vis

    pathway = KGML_parser.read(f.read().decode('UTF-8'))

    kgml_map = KGMLCanvas(pathway, show_maps=True)

    # Let's use some arbitrary colours for the orthologs
    cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5, jitter=0.03)
    # Loop over the orthologs in the pathway, and change the
    # background colour
    orthologs = [e for e in pathway.orthologs]
    for o in orthologs:
        match = False
        if 'K00163' in o.name:
            print('##################################')
        ko_temp_list = set([i.rstrip() for i in o.name.split('ko:')])
        if len(ko_temp_list.intersection(set(ko2freq.keys()))) > 0:

            ko_keep = []
            for ko in ko_temp_list:
                if ko in ko2freq:
                    ko_keep.append(ko)
                if ko in ko_list:
                    match = True
            o.name = 'ko:' + ' ko:'.join(ko_keep)
            total = sum([
                int(ko2freq[i])
                for i in ko_temp_list.intersection(set(ko2freq.keys()))
            ])

            for g in o.graphics:
                if match:
                    g.bgcolor = rgb2hex(m2.to_rgba(float(total)))
                else:
                    #print 'no match!!!!'
                    #print ko_temp_list
                    #print ko2freq.keys()
                    #print 'TOTAL:', total
                    g.bgcolor = rgb2hex(m.to_rgba(float(total)))
            o.name = "%s (%s)" % (o.name.split('ko:')[0], total)
        #else:
        #    for g in o.graphics:
        #        g.bgcolor = '#FFFFFF'

    # Default settings are for the KGML elements only

    # We need to use the image map, and turn off the KGML elements, to see
    # only the .png base map. We could have set these values on canvas
    # instantiation
    kgml_map.import_imagemap = True
    kgml_map.show_maps = True
    kgml_map.show_orthologs = True
    kgml_map.draw_relations = False
    kgml_map.show_compounds = False
    kgml_map.show_genes = False
    kgml_map.show_compounds = False
    kgml_map.show_genes = False
    kgml_map.draw(outpath)
    '''
    print 'DIRLISAT:', dir(pathway)
    maps = [m for m in pathway.maps]
    for map in maps:
        for g in map.graphics:
            print g.name
    '''

    #print re.sub('pdf', 'svg', outpath)
    shell_command.shell_command(
        'inkscape %s --export-plain-svg=%s' %
        (outpath, re.sub('pdf', 'svg', outpath)))  # 'pdf2svg %s %s all'
    t = edit_svg_map("%s" % re.sub('pdf', 'svg', outpath),
                     ko2freq.keys(),
                     biodb,
                     map_id,
                     taxon_id=taxon_id)
    #print "%s" % re.sub('pdf', 'svg', outpath)
    t.write("%s" % re.sub('pdf', 'svg', outpath))
def gatherDetails(enterPathway,folderName,useCO,CO_values):
    #check if the directories exist, one for pathway files
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    #else:
        #raise ValueError('Be careful, this folder already exists')
                   
    #only one pathway at a time
    setKeep = 1
    try:
        kegg_get(enterPathway).read()
    except:
        #use the ko map if there is nothing species specific...this can also fail...
        usePathway = 'ko' + enterPathway[3:8]
        setKeep = 0
        try:
            kegg_get(usePathway).read()
        except:
            pass
        
    if setKeep:
        usePathway = enterPathway

    #get the compounds and genes for this pathway     
    genes = getKfrom_ko(usePathway)
    compounds = getCfrom_ko(usePathway)
    
    #figure out which ones I have data for...
    setG = set(genes)
    setC = set(compounds)
    setT = set(useCO)
    intCompounds = setC.intersection(setT)
        
    ## plot the pathway map for this pathway, get details from KEGG for plotting (%must be at least 4 colors)
    useColors = pal.colorbrewer.diverging.PuOr_4.hex_colors
    #useColors = pal.colorbrewer.diverging.RdYlBu_11.hex_colors
    
    #set the color of the mtab based on its value, only scale the values from this particular pathway
    useCOsubset = CO_values.loc[intCompounds]
    cmin = useCOsubset.min() #find min and max...ignore NaN and inf for the moment
    cmax = useCOsubset.replace([np.inf],np.nan).dropna(how = 'all').max()
    
    size = 20 #increase the size of the compounds in the plots
        
    #can have all zeros...
    if sum(useCOsubset.dropna())==0:
        pass
        #print('No measured metabolites in pathway ' + usePathway)
    elif len(useCOsubset.value_counts())==1:
        #only two color options: yes/no
        dummy = useCOsubset.copy(deep = True)
        dummy.replace([np.inf],np.nan,inplace = True)
        for idx,item in enumerate(useCOsubset):
            if np.isnan(item):
                useCOsubset.iloc[idx] = int(0)
            else:
                useCOsubset.iloc[idx] = int(1) 
        
        #go get the pathway information and customize the plot
        pathway = KGML_parser.read(kegg_get(usePathway, "kgml")) #no choice in gene color: green

        # Change the colors of compounds
        for element in pathway.compounds:
            for graphic in element.graphics:
                tc = element.name[4:10] #skip over the 'cpd:'
                if (tc in intCompounds):
                    #in the pathway, set the color
                    tempColor = useCOsubset.loc[tc]
                    graphic.bgcolor = useColors[int(tempColor)] 
                    graphic.width = size
                    graphic.height = size

        canvas = KGMLCanvas(pathway, import_imagemap=True)
        pdfName = 'mapWithColors_' + str(usePathway) + '.pdf'
        canvas.draw(folderName + '/' + pdfName)
        pdfName = None #empty it in case that is where I am having issues         
        
    else:
        dummy = useCOsubset.copy(deep = True)
        dummy.replace([np.inf],np.nan,inplace = True)
        for idx,item in enumerate(useCOsubset):
            if np.isnan(item):
                useCOsubset.iloc[idx] = 0
            elif np.isinf(item):
                useCOsubset.iloc[idx] = 10*cmax #make inf 10x the biggest value

        #now, find cmax again...use that downstream
        cmax = useCOsubset.replace([np.inf],np.nan).dropna(how = 'all').max()

        #use histogram to make the bins (complete hack)
        a,bin_edges = np.histogram(useCOsubset,bins = len(useColors)-3,range = (cmin,cmax))
        #now...put zero at beginning and inf at end
        #BUT - can actually have cases with values for all metabolites (novel concept)
        try:
            nz = useCOsubset.value_counts()[0] #count the number of zeros
            a = np.insert(a,0,nz)
            bin_edges = np.insert(bin_edges,0,0)
        except:
            pass
            
        try:
            nm = useCOsubset.value_counts()[cmax]
            a = np.append(a,nm)
            bin_edges = np.append(bin_edges,cmax)
        except:
            pass

        #then find the index for each number...this will be the index into useColors
        useIdx = np.digitize(useCOsubset,bin_edges)
        color_df = pd.DataFrame({'mtab': useCOsubset,'idx':useIdx})

        #go get the pathway information and customize the plot
        pathway = KGML_parser.read(kegg_get(usePathway, "kgml")) #no choice in gene color: green

        # Change the colors of compounds
        for element in pathway.compounds:
            for graphic in element.graphics:
                tc = element.name[4:10] #skip over the 'cpd:'
                if (tc in intCompounds):
                    #in the pathway, set the color
                    tempColor = color_df.loc[tc,'idx']
                    graphic.bgcolor = useColors[int(tempColor)-1] 
                    graphic.width = size
                    graphic.height = size

        canvas = KGMLCanvas(pathway, import_imagemap=True)
        pdfName = 'mapWithColors_' + str(usePathway) + '.pdf'
        #Tracer()()
        canvas.draw(folderName + '/' + pdfName)
        pdfName = None #empty it in case that is where I am having issues
예제 #15
0
# process all found kegg pathways
for k in kegg:
    print("Processing: {}".format(k))
    stats[k] = defaultdict(int)
    processedIDs = set()
    # load current pathway
    isRef = True
    try:
        pathway = KGML_parser.read(kegg_get("{}{}".format(keggmap,k), "kgml"))
    except:
        print("{} not a reference pathway.".format(k))
        pathway = KGML_parser.read(kegg_get("ko{}".format(k), "kgml"))
        isRef = False
        
        
    canvas = KGMLCanvas(pathway, import_imagemap=True)
    if isRef:
        for element in list(pathway.genes):
            f = kegg_find(keggmap, element.name)
            results = f.readline() 
            g = results.split("\t")[0]#[len(keggmap)+1:]
            
            for l in kegg_get(g).readlines():
                if "ORTHOLOGY" in l:
                    match = pattern.findall(l) # EC number
                    ortho= l[12:18]
                    #print(match) 
                    if len(match) == 0:
                        print("{} has no EC".format(ortho))
                        eName = set([x.split(":")[1] for x in element.name.split(" ")])
                        if ortho not in processedIDs:
예제 #16
0
                    [x for x in keggID[1:] if x not in kegg[keggID[0]]])
                org = proteinMapping[currentGene]
                for ec in keggID[1:]:
                    if org not in ecToGeneOrg:
                        ecToGeneOrg[org] = defaultdict(list)
                    if currentGene not in ecToGeneOrg[org][ec]:
                        ecToGeneOrg[org][ec].append(currentGene)

# process all found kegg pathways
for k in kegg:
    print("Processing: {}".format(k))
    stats[k] = defaultdict(int)
    processedIDs = set()
    # load current pathway
    pathway = KGML_parser.read(kegg_get("ko{}".format(k), "kgml"))
    canvas = KGMLCanvas(pathway, import_imagemap=True)

    # get information on EC numbers in kegg pathway
    for ec in kegg[k]:
        print(" EC: {}".format(ec))
        if True:
            foundOrtho = False
            # query KEGG
            for ecInfo in kegg_get("ec:{}".format(ec)):
                ecInfoLabel = ecInfo[:12]
                if "ORTHOLOGY" in ecInfoLabel:
                    foundOrtho = True
                    KOToEC[ecInfo[12:18]].append(ec)
#                    KOToGene[ecInfo[12:18]].extend(ecToGene[ec])
                else:
                    foundOrtho = foundOrtho and len(ecInfoLabel.strip()) == 0
def gatherDetails(makeNclusters,trimPath,forRelatedness,folderName,CO_fromMATLAB,KO_Norm2Mean,Insitu_TPM_DIA,Insitu_TPM_DIN,Insitu_TPM_Oth):
    colLabel = ['nCpds','nGenes'] #starting with this is easiest - makes one list, no need to flatten

    for item in range(makeNclusters):
        colLabel.append('Km' + str(item) + '_cpd')
        colLabel.append('Km' + str(item) + '_gene')

    gatherCounts = pd.DataFrame(0, index = trimPath, columns = colLabel)

    #setup the strings to match first
    rnString = re.compile('(?:[rn:R])(\d+)$') #will return R00190
    cpdString = re.compile('(?:[cpd:C])(\d+)$') #will return C00190

    size = 20 #turns out I can increase the size of the compounds in the plots

    for kn in range(makeNclusters):
        fullSet = set(forRelatedness.KEGG)
        oneK = forRelatedness[forRelatedness.kmeans == kn] #get gene & transcript information for one Kmeans group
        getKm = 'Km' + str(kn)

        #check if the directories exist, one for pathway files
        directoryPDF = folderName + str(kn) + '/pathway_files'
        if not os.path.exists(directoryPDF):
            os.makedirs(directoryPDF)
        else:
            raise ValueError('Krista - be careful, this folder already exists')

        #check if the directories exist, one for reaction files
        directoryPNG = folderName + str(kn) + '/reaction_files'
        if not os.path.exists(directoryPNG):
            os.makedirs(directoryPNG) 
        else:
            raise ValueError('Krista - be careful, this folder already exists')
                       
        #check if the directories exist, one for species 
        directorySpecies = folderName + str(kn) + '/species_files'
        if not os.path.exists(directorySpecies):
            os.makedirs(directorySpecies) 
        else:
            raise ValueError('Krista - be careful, this folder already exists')
                    
        for item in trimPath: #searching within one pathway at a time
            plotPathway = [] #gather up yes/no and will only plot if have linked genes/mtabs    
            genes = getKfrom_ko(item)
            compounds = getCfrom_ko(item)
            gatherCounts.loc[item,'nCpds'] = len(compounds)
            gatherCounts.loc[item,'nGenes'] = len(genes)     
            #have to track genes and compounds differently for the biopython plotting later on 
            setG = set(genes)
            setC = set(compounds)
            setB = set(oneK.KEGG)
            intGenes = setG.intersection(setB)
            intCompounds = setC.intersection(setB)
            gatherCounts.loc[item,(getKm + '_gene')] = len(intGenes)
            gatherCounts.loc[item,(getKm + '_cpd')] = len(intCompounds)
            for gen in intGenes: #go through each gene...one at a time
                rnList = kegg_link('reaction',gen).read() #get the list of reactions for that gene
                #can have cases where there is a gene and no reaction (K02906 for example). This returns rnList = '\n'
                #since this is not actually empty...need a few way to filter those out
                test = '\n'
                if test != rnList:
                    for line in rnList.rstrip().split('\n'):
                        countCpd = []
                        countGene = []
                        m = rnString.search(line) #get the reaction number
                        cpdList = kegg_link('cpd',m.group(0)).read() #now go get the compounds for that reaction
                        #can have no compounds in a reaction (only glycans, begin with G, nothing I have matched)
                        if len(cpdList) > 1: #will be true if cpdList includes compounds
                            for line2 in cpdList.rstrip().split('\n'):
                                m2 = cpdString.search(line2).group(0)
                                #now that I have a compound, check if it is in intCompounds
                                if m2 in intCompounds:
                                    countCpd.append(m2) 
                                    countGene.append(gen)
                                    plotPathway.append('yes')
                        ##Now, plot the PNG files (one for each reaction within a pathway)
                        if len(countCpd) > 0:
                            dayList = ['S1','S2','S3','S4','S5']
                            kData = pd.DataFrame(columns = dayList)
                            for k in set(countGene):
                                kData = kData.append(oneK.ix[k,dayList])
                            cData = pd.DataFrame(columns = dayList)
                            for co in set(countCpd):
                                #convert CO to RI, can have multiple options
                                j = findRInumber(oneK,co)
                                cData = cData.append(oneK.loc[j,dayList])
                            fig,ax = plt.subplots(1)
                            cData.T.plot(color = 'k',ax=ax)
                            kData.T.plot(color = 'r',ax=ax)
                            handles, labels = ax.get_legend_handles_labels()
                            #convert the RI numbers to COnumbers for the figure
                            for ia, a in enumerate(labels):
                                #add compound/gene name to the legend
                                if a[0]== 'R':
                                    tLabel = convertRItoCO(CO_fromMATLAB,a)
                                    fn = kegg_list(tLabel).read()                          
                                    labels[ia] = fn
                                elif a[0] == 'K':
                                    fn = kegg_list(a).read()
                                    labels[ia] = fn
                            ax.legend(handles, labels, bbox_to_anchor = ([-1, 0.5]))
                            fig.suptitle('pathway ' + item + ', Kmeans grp ' + str(kn))
                            pngName = 'pathway' + item + '_' + m.group(0) + '.png'
                            fig.savefig(directoryPNG + '/' + pngName, bbox_inches = 'tight')
                            pngName = None #empty it in case that is where I am having issues
                            plt.close()
            if len(plotPathway)>0:
                ## plot the pathway map for this pathway, get details from KEGG for plotting
                useColors = pal.colorbrewer.qualitative.Set1_4.hex_colors
                useColors.insert(0,'#f7f7f7') ## insert white at beginning
                # order of colors: white, red, blue,green,purple
                sd = 0 #not in dataset
                sk = 1 #in K means group and pathway
                sa = 2 #in pathway, in any K means (for genes, bc overlap in numbers)
                sn = 3 #in pathway, not in K means group (compounds only)               
                su = 4 #unconnected gene or compound
                line1 = useColors[sd] + ', not in dataset' + '\n'
                line2 = useColors[sk] + ', in K means group and pathway' + '\n'
                line3 = useColors[sa] + ', #in pathway, in any K means (for genes, bc overlap in numbers)' +'\n'
                line4 = useColors[sn] +  ', #in pathway, not in K means group (compounds only)' + '\n'               
                line5 = useColors[su] + ', #unconnected gene or compound' + '\n'
                file = open("readme_colorsInPathways.txt", "w")
                file.write(line1 + line2 + line3 + line4 + line5)
                file.close()
                
                pathway = KGML_parser.read(kegg_get(item, "kgml"))
                for element in pathway.orthologs:
                    #print element.name
                    for graphic in element.graphics:
                        tg = element.name[3:9] #skip over the 'ko:'
                        if (tg in intGenes):
                            #in the pathway AND in the set for this particular K means group
                            graphic.bgcolor = useColors[sk] #
                            
                            #if this is something in the pathway, plot up the species for the K number
                            if tg in Insitu_TPM_DIA.index.tolist():
                                Dk=Insitu_TPM_DIA.loc[tg]
                            else: 
                                Dk = 0/Insitu_TPM_DIA.iloc[0] #make an empty frame
                            if tg in Insitu_TPM_DIN.index.tolist():
                                Nk=Insitu_TPM_DIN.loc[tg]
                            else:
                                Nk = 0/Insitu_TPM_DIN.iloc[0]
                            if tg in Insitu_TPM_Oth.index.tolist():
                                Ok=Insitu_TPM_Oth.loc[tg]
                            else:
                                Ok = 0/Insitu_TPM_Oth.iloc[0]
                            fig,ax=plt.subplots(1)
                            ax.stackplot(range(5), Dk, Nk, Ok, colors=pal.colorbrewer.qualitative.Set3_6_r.hex_colors, lw=0)
                            ax.set_xticks(range(5))
                            ax.set_xticklabels([1,2,3,4,5])
                            ax.set_ylabel('In situ TPM')
                            plt.title(tg + ', lt orange=diatoms, blue=dinos, dk orange=other')
                            fig.savefig(directorySpecies + '/' + tg + '_species.png',bbox_inches='tight')
                            plt.close()
                        elif (tg in fullSet) and (tg in genes) and (tg not in intGenes):
                            #in the pathway AND in the set of genes from RI, allow any Kmeans group for genes
                            graphic.bgcolor = useColors[sa] #
                        elif (tg not in fullSet) and (tg in genes) and (tg not in KO_Norm2Mean.index.tolist()):
                            #in the pathway, but *not* in anything from the RI samples
                            graphic.bgcolor = useColors[sd] #
                        elif (tg not in fullSet) and (tg in genes) and (tg in KO_Norm2Mean.index.tolist()): 
                            #an unconnected gene in the RI data
                            graphic.bgcolor = useColors[su] #
                # Change the colours of compounds (mostly same as genes
                for element in pathway.compounds:
                    for graphic in element.graphics:
                        tc = element.name[4:10] #skip over the 'cpd:'
                        if (tc in intCompounds):
                            #in the pathway AND in the set for this particular K means group
                            graphic.bgcolor = useColors[sk] #
                            graphic.width = size
                            graphic.height = size
                        elif (tc in fullSet) and (tc in compounds) and (tc not in intCompounds):
                            #in the pathway AND in the set of compounds from RI, but *not* in this Kmeans group
                            graphic.bgcolor = useColors[sn] #
                            graphic.width = size
                            graphic.height = size
                        elif (tc not in fullSet) and (tc in compounds) and (tc not in CO_fromMATLAB.cNumber.values):
                            #in the pathway, but *not* in anything from the RI samples
                            graphic.bgcolor = useColors[sd] #  
                        elif (tc not in fullSet) and (tc in compounds) and (tc in CO_fromMATLAB.cNumber.values): #seems like a hack
                            #unconnected compound in the RI data
                            graphic.bgcolor = useColors[su] #
                            graphic.width = size
                            graphic.height = size
                canvas = KGMLCanvas(pathway, import_imagemap=True)
                pdfName = 'mapWithColors_' + str(item) + '.pdf'
                canvas.draw(directoryPDF + '/' + pdfName)
                pdfName = None #empty it in case that is where I am having issues
    #stick the pathway information into gatherCounts before I export...
    #want to export gatherCounts, with the added pathway name as a new column
    gatherCounts['pathwayInfo'] = ''
    gatherCounts['pathwayGroup_A'] = ''
    gatherCounts['pathwayGroup_B'] = ''
    gatherCounts['pathwayGroup_C'] = ''
    #go read in the file from KEGG
    D = glob.glob('br08901.keg') #from http://www.genome.jp/kegg-bin/get_htext?br08901.keg; 3/15/2016
    allBRITE=[]
    for idx,nof in enumerate(D):
        allBRITE = ReadBRITEfile(nof) 

    #put the pathway name and group into the data frame before exporting it
    for item in gatherCounts.index:
        #if this error appears: IndexError: index 0 is out of bounds for axis 0 with size 0
        #KEGG has updated a pathway, but not the BRITE file (see below for work around)
        pathstr = kegg_list(item).read()
        #this next line splits the string at the '\t', then keeps the piece at index = 1, and strips off the '\n'
        gatherCounts.loc[item,('pathwayInfo')] = pathstr.split('\t')[1].rstrip()
        t = allBRITE.loc[allBRITE['map']==item[2:]]  
        #put in a check to see if t.empty ...will be empty if KEGG updated pathway and not BRITE file
        if t.empty is False: 
            gatherCounts.set_value(item,'pathwayGroup_A',t['A'].values[0])
            gatherCounts.set_value(item,'pathwayGroup_B',t['B'].values[0])
            gatherCounts.set_value(item,'pathwayGroup_C',t['C'].values[0])
    
    return gatherCounts
예제 #18
0
    def pathway_pdf(self, filename = "", imagemap = True, orthologs = True, compounds = True, maps = True, reactions = True):
        '''
        Prints current pathway to PDF file
        :param filename: (str) - PDF filename
        :param imagemap: (bol) - Print imagemap
        :param orthologs: (bol) - Print orthologs
        :param compounds: (bol) - Print compounds
        :param maps: (bol) - Print maps
        :param reactions: (bol) - Print reactions ???
        :return: PDF file with current pathway
        '''
        #TODO Verificar o parametro reactions
        pathway_pdf = KGMLCanvas(self.pathway)
        pathway_pdf.import_imagemap=imagemap
        pathway_pdf.label_orthologs = orthologs
        pathway_pdf.label_compounds = compounds
        pathway_pdf.label_maps = maps
        pathway_pdf.label_reaction_entries = reactions

        if filename == "":
            pathway_pdf.draw(str(self.pathway_ID+".pdf"))
        else:
            pathway_pdf.draw(str(filename+".pdf"))