gwaslab 3.4.47__py3-none-any.whl → 3.4.49__py3-none-any.whl

gwaslab/util_ex_ldsc.py

@@ -304,9 +304,16 @@ def _estimate_h2_by_ldsc(insumstats, log, verbose=True, munge=False, munge_args=
     log.write(" -LDSC log:", verbose=verbose)
     summary = estimate_h2(sumstats, default_args, log)
     
+    results_table = None
+    if type(summary) is tuple:
+        results_table = summary[1]
+        summary = summary[0]
+        log.write(" -Coefficient results have been stored in .ldsc_h2_results", verbose=verbose)
+        
+
     log.write(" -Results have been stored in .ldsc_h2", verbose=verbose)
     finished(log=log,verbose=verbose,end_line=_end_line)
-    return parse_ldsc_summary(summary)
+    return parse_ldsc_summary(summary), results_table
 
 
 ####################################################################################################################

gwaslab/util_ex_run_clumping.py

@@ -11,7 +11,7 @@ from gwaslab.g_version import _checking_plink_version
 def _clump(insumstats, vcf=None, scaled=False, out="clumping_plink2", 
            p="P",mlog10p="MLOG10P", overwrite=False, study=None, bfile=None, 
            n_cores=1, memory=None, chrom=None, clump_p1=5e-8, clump_p2=5e-8, clump_r2=0.01, clump_kb=250,
-           log=Log(),verbose=True):
+           log=Log(),verbose=True,plink="plink",plink2="plink2"):
     ##start function with col checking##########################################################
     _start_line = "perfrom clumping"
     _end_line = "clumping"
@@ -111,7 +111,7 @@ def _clump(insumstats, vcf=None, scaled=False, out="clumping_plink2",
             bfile_to_use = bfile
 
         log.write(" -Performing clumping for CHR {}...".format(i),verbose=verbose)
-        log = _checking_plink_version(v=2, log=log)
+        log = _checking_plink_version(plink2=plink2, log=log)
         if memory is not None:
             memory_flag = "--memory {}".format(memory)
         
@@ -123,7 +123,7 @@ def _clump(insumstats, vcf=None, scaled=False, out="clumping_plink2",
         if scaled == True:
             # clumping using LOG10P
             script = """
-            plink2 \
+            {} \
                 {}\
                 --chr {} \
                 --clump {} \
@@ -136,11 +136,11 @@ def _clump(insumstats, vcf=None, scaled=False, out="clumping_plink2",
                 --clump-kb {} \
                 --threads {} {}\
                 --out {}
-            """.format(file_flag, chrom, clump, mlog10p,clump_log10_p1, clump_log10_p2, clump_r2, clump_kb, n_cores, memory_flag if memory is not None else "", out_single_chr)    
+            """.format(plink2, file_flag, chrom, clump, mlog10p,clump_log10_p1, clump_log10_p2, clump_r2, clump_kb, n_cores, memory_flag if memory is not None else "", out_single_chr)    
         else:
             # clumping using P
             script = """
-            plink2 \
+            {} \
                 {}\
                 --chr {} \
                 --clump {} \
@@ -152,7 +152,7 @@ def _clump(insumstats, vcf=None, scaled=False, out="clumping_plink2",
                 --clump-kb {} \
                 --threads {} {}\
                 --out {}
-            """.format(file_flag, chrom, clump, p, clump_p1, clump_p2, clump_r2, clump_kb, n_cores,memory_flag if memory is not None else "", out_single_chr)
+            """.format(plink2,file_flag, chrom, clump, p, clump_p1, clump_p2, clump_r2, clump_kb, n_cores,memory_flag if memory is not None else "", out_single_chr)
         
         try:
             output = subprocess.check_output(script, stderr=subprocess.STDOUT, shell=True,text=True)

gwaslab/util_in_fill_data.py

@@ -184,7 +184,8 @@ def fill_mlog10p(sumstats,log,verbose=True,filled_count=0):
     else:
         return 0,filled_count
     return 1,filled_count
-def fill_extreme_mlog10p(sumstats,log,verbose=True,filled_count=0):
+
+def fill_extreme_mlog10p(sumstats,df,log,verbose=True,filled_count=0):
     # ref: https://stackoverflow.com/questions/46416027/how-to-compute-p-values-from-z-scores-in-r-when-the-z-score-is-large-pvalue-muc/46416222#46416222
     if "Z" in sumstats.columns:
         # P -> MLOG10P
@@ -198,6 +199,10 @@ def fill_extreme_mlog10p(sumstats,log,verbose=True,filled_count=0):
         log.write("  - Filling MLOG10P using Z column...", verbose=verbose)
         sumstats = fill_extreme_mlog10(sumstats, "Z")
         filled_count +=1
+    elif "CHISQ" in sumstats.columns and "DOF" in sumstats.columns:
+        log.write("  - Filling MLOG10P using CHISQ and DOF column...", verbose=verbose)
+        sumstats = fill_extreme_mlog10_chisq(sumstats, "CHISQ", df)
+        filled_count +=1
     else:
         return 0,filled_count
     return 1,filled_count
@@ -223,6 +228,19 @@ def fill_extreme_mlog10(sumstats, z):
     sumstats["P_EXPONENT"]= exponent
     return sumstats
 
+def fill_extreme_mlog10_chisq(sumstats, chisq, df):
+    #https://stackoverflow.com/a/46416222/199475
+    log_pvalue = ss.chi2.logsf(sumstats[chisq], sumstats[df])
+
+    log10_pvalue = log_pvalue/np.log(10)
+    
+    mantissa = 10**(log10_pvalue %1)
+    exponent = log10_pvalue // 1
+    sumstats["MLOG10P"] = -log10_pvalue
+    sumstats["P_MANTISSA"]= mantissa
+    sumstats["P_EXPONENT"]= exponent
+    return sumstats
+
 ####################################################################################################################
 def fill_iteratively(sumstats,raw_to_fill,log,only_sig,df,extreme,verbose,sig_level):
     to_fill = raw_to_fill.copy()
@@ -260,7 +278,7 @@ def fill_iteratively(sumstats,raw_to_fill,log,only_sig,df,extreme,verbose,sig_le
     # p to -log10(P)  ###############################################################################################
         if "MLOG10P" in to_fill:
             if extreme==True:
-                status,filled_count = fill_extreme_mlog10p(sumstats,log,verbose=verbose,filled_count=filled_count)
+                status,filled_count = fill_extreme_mlog10p(sumstats,df, log,verbose=verbose,filled_count=filled_count)
                 filled_count +=1
             elif "P" not in sumstats.columns:
                 fill_p(sumstats,log,verbose=verbose)

gwaslab/viz_aux_annotate_plot.py

@@ -38,6 +38,7 @@ def annotate_single(
     region,
     region_anno_bbox_args,
     skip,
+    anno_height=1,
     amode="int",
     snpid="SNPID",
     chrom="CHR",
@@ -131,7 +132,7 @@ def annotate_single(
             
             #xy=(row["i"],row["scaled_P"]+0.2)
             xy=(row["i"],row["scaled_P"]+0.01*maxy)
-            xytext=(last_pos,1.15*maxy*arm_scale)
+            xytext=(last_pos,1.15*maxy*arm_scale*anno_height)
             
             if anno_fixed_arm_length is not None:
                 armB_length_in_point = anno_fixed_arm_length

gwaslab/viz_aux_quickfix.py

@@ -286,8 +286,9 @@ def _cut(series, mode,cutfactor,cut,skip, ylabels, cut_log, verbose, lines_to_pl
     log.write(" -Converting data above cut line...",verbose=verbose)
     if ylabels is not None:
         ylabels = pd.Series(ylabels)
-    maxy = series.max()
     series = series.copy()
+    
+    maxy = series.max()
     if "b" not in mode:
         log.write(" -Maximum -log10(P) value is "+str(maxy) +" .", verbose=verbose)
     elif "b" in mode:

gwaslab/viz_plot_compare_effect.py

@@ -77,8 +77,10 @@ def compare_effect(path1,
         scaled2 = True
     if is_q_mc=="fdr" or is_q_mc=="bon":
         is_q = True
-    else:
-        raise ValueError("Please select either fdr or bon for is_q_mc.")
+    
+    if is_q == True:
+        if is_q_mc not in [False,"fdr","bon","non"]:
+            raise ValueError("Please select either fdr or bon or non for is_q_mc.")
     if save_args is None:
         save_args = {"dpi":300,"facecolor":"white"}
     if reg_box is None:

gwaslab/viz_plot_miamiplot2.py

@@ -247,7 +247,10 @@ def plot_miami2(
         plt.subplots_adjust(hspace=region_hspace)    
     else:
         fig, ax1, ax5 = figax
-    
+
+    #if same_ylim==True:
+        #maxy = merged_sumstats[["scaled_P_1","scaled_P_2"]].max().max()
+
     log.write("Start to create Manhattan plot for sumstats1...", verbose=verbose)
     fig,log = mqqplot(merged_sumstats,
                       chrom="CHR",
@@ -284,16 +287,14 @@ def plot_miami2(
                       _if_quick_qc=False,
                      **mqq_args2)
     log.write("Finished creating Manhattan plot for sumstats2".format(_get_version()), verbose=verbose)
+    
 
-    if same_ylim==True:
-        ylim1_converted = ax1.get_ylim()
-        ylim2_converted = ax5.get_ylim()
-        if ylim1_converted > ylim2_converted:
-            ax5.set_ylim(ylim1_converted)
-        else:
-            ax1.set_ylim(ylim2_converted)
     #####################################################################################################################    
-    
+    ax1l, ax1r = ax5.get_xlim()
+    ax5l, ax5r = ax1.get_xlim()
+    ax1.set_xlim([min(ax1l,ax5l), max(ax1r,ax5r)])
+    ax5.set_xlim([min(ax1l,ax5l), max(ax1r,ax5r)])
+    #####################################################################################################################
     ax5.set_xlabel("")
     #ax5.set_xticks(chrom_df)
     ax5.set_xticklabels([])

gwaslab/viz_plot_mqqplot.py

@@ -141,6 +141,7 @@ def mqqplot(insumstats,
           anno_max_iter=100,
           arm_offset=50,
           arm_scale=1,
+          anno_height=1,
           arm_scale_d=None,
           cut=0,
           skip=0,
@@ -180,6 +181,7 @@ def mqqplot(insumstats,
           xpad=None,
           xpadl=None,
           xpadr=None,
+          xtight=False,
           chrpad=0.03, 
           drop_chr_start=False,
           title =None,
@@ -552,7 +554,8 @@ def mqqplot(insumstats,
                                                                         cut_log = cut_log,
                                                                         verbose =verbose, 
                                                                         lines_to_plot=lines_to_plot,
-                                                                        log = log)
+                                                                        log = log
+                                                                        )
     except:
         log.warning("No valid data! Please check the input.")
         return None
@@ -596,19 +599,23 @@ def mqqplot(insumstats,
             sumstats.loc[sumstats["scaled_P"]>-np.log10(sig_level_plot),"s"]=4
         sumstats["chr_hue"]=sumstats[chrom].astype("string")
 
-        if vcf_path is not None:
+        if "r" in mode:
+            if vcf_path is None:
+                sumstats["LD"]=100
+                sumstats["SHAPE"]=1
             sumstats["chr_hue"]=sumstats["LD"]
+                
         ## default seetings
         
         palette = sns.color_palette(colors,n_colors=sumstats[chrom].nunique())  
-        
 
         legend = None
         style=None
         linewidth=0
         edgecolor="black"
         # if regional plot assign colors
-        if vcf_path is not None:
+        if "r" in mode:
+            #if vcf_path is not None: 
             legend=None
             linewidth=1
             if len(region_ref) == 1:
@@ -631,10 +638,9 @@ def mqqplot(insumstats,
                         palette[(i+1)*100 + j ] = hex_color
 
                 edgecolor="none"
-                scatter_args["markers"]= region_marker_shapes[:len(region_ref)]
+                scatter_args["markers"]= {(i+1):m for i,m in enumerate(region_marker_shapes[:len(region_ref)])}
                 style="SHAPE"
-                
-            
+
             
         ## if highlight 
         highlight_i = pd.DataFrame()
@@ -977,6 +983,7 @@ def mqqplot(insumstats,
                                 region=region,
                                 region_anno_bbox_args=region_anno_bbox_args,
                                 skip=skip,
+                                anno_height=anno_height,
                                 snpid=snpid,
                                 chrom=chrom,
                                 pos=pos,
@@ -1040,7 +1047,7 @@ def mqqplot(insumstats,
         if "qq" in mode:
             ax2.set_ylim(ylim)
     
-    ax1 = _add_pad_to_x_axis(ax1, xpad, xpadl, xpadr, sumstats)
+    ax1 = _add_pad_to_x_axis(ax1, xpad, xpadl, xpadr, sumstats, pos, chrpad, xtight, log = log, verbose=verbose)
 
     # Titles 
     if title and anno and len(to_annotate)>0:
@@ -1065,20 +1072,34 @@ def mqqplot(insumstats,
 
 
 
-def _add_pad_to_x_axis(ax1, xpad, xpadl, xpadr, sumstats):
+def _add_pad_to_x_axis(ax1, xpad, xpadl, xpadr, sumstats, pos, chrpad, xtight, log, verbose):
     
-    if ax1 is not None:
-        xmin, xmax = ax1.get_xlim() 
-        
-        if xpad is not None:
-            pad = xpad* sumstats["i"].max()
-            ax1.set_xlim([xmin - pad, xmin + pad])
-        if xpadl is not None:
-            pad = xpadl* sumstats["i"].max()
-            ax1.set_xlim([xmin - pad,xmax])
-        if xpadr is not None:
-            pad = xpadr* sumstats["i"].max()
-            ax1.set_xlim([xmin, xmax + pad])
+    if xtight==True:
+        log.write(" -Adjusting X padding on both side : tight mode", verbose=verbose)
+        xmax = sumstats["i"].max()
+        xmin=  sumstats["i"].min()
+        ax1.set_xlim([xmin, xmax])
+    
+    else:
+        chrpad_to_remove = sumstats[pos].max()*chrpad
+        if ax1 is not None:
+            xmin, xmax = ax1.get_xlim() 
+            length = xmax - xmin
+            
+            if xpad is not None:
+                log.write(" -Adjusting X padding on both side: {}".format(xpad), verbose=verbose)
+                pad = xpad* length #sumstats["i"].max()
+                ax1.set_xlim([xmin - pad + chrpad_to_remove, xmax + pad - chrpad_to_remove])
+            if xpad is None and xpadl is not None:
+                log.write(" -Adjusting X padding on left side: {}".format(xpadl), verbose=verbose)
+                xmin, xmax = ax1.get_xlim() 
+                pad = xpadl*length # sumstats["i"].max()
+                ax1.set_xlim([xmin - pad + chrpad_to_remove ,xmax])
+            if xpad is None and xpadr is not None:
+                log.write(" -Adjusting X padding on right side: {}".format(xpadr), verbose=verbose)
+                xmin, xmax = ax1.get_xlim() 
+                pad = xpadr*length # sumstats["i"].max()
+                ax1.set_xlim([xmin, xmax + pad - chrpad_to_remove])
 
     return ax1
 

gwaslab/viz_plot_regional2.py

@@ -5,6 +5,7 @@ import matplotlib.patches as patches
 import seaborn as sns
 import numpy as np
 import copy
+import re
 import scipy as sp
 from pyensembl import EnsemblRelease
 from allel import GenotypeArray
@@ -96,19 +97,20 @@ def _plot_regional(
                                         marker_size= marker_size,
                                         region_marker_shapes=region_marker_shapes,
                                         log=log,verbose=verbose)
-            if lead_id_single is not None:
-                lead_ids.append(lead_id_single)       
+            #if lead_id_single is not None:
+            lead_ids.append(lead_id_single)       
         
         # update region_ref to variant rsID or variantID / skip NAs
         new_region_ref = []
         for i in range(len(lead_ids)):
             if lead_ids[i] is None:
+                new_region_ref.append(region_ref[i])
                 continue
             if region_ref[i] is None:
-                if "rsID" in sumstats.columns:
-                    new_name = sumstats.loc[lead_ids[i],"rsID"]
-                elif "SNPID" in sumstats.columns:
+                if "SNPID" in sumstats.columns:
                     new_name = sumstats.loc[lead_ids[i],"SNPID"]
+                elif "rsID" in sumstats.columns:
+                    new_name = sumstats.loc[lead_ids[i],"rsID"]
                 else:
                     new_name = "chr{}:{}".format(sumstats.loc[lead_ids[i],"CHR"] , sumstats.loc[lead_ids[i],"POS"])
                 new_region_ref.append(new_name)
@@ -162,7 +164,6 @@ def _plot_regional(
         lead_snp_ys = []
         lead_snp_is = []
         lead_snp_is_colors = []
-        
         for i,lead_id_single in enumerate(lead_ids):
             if lead_id_single is not None:
                 lead_snp_ys.append(sumstats.loc[lead_id_single,"scaled_P"] )
@@ -258,11 +259,35 @@ def _get_lead_id(sumstats=None, region_ref=None, log=None, verbose=True):
     if type(lead_id) is list:
         if len(lead_id)>0:
             lead_id = int(lead_id[0])
-    
+
     if region_ref_to_check is not None:
         if type(lead_id) is list:
             if len(lead_id)==0 :
-                log.warning("{} not found.. Skipping..".format(region_ref_to_check))
+                #try:
+                matched_snpid = re.match("(chr)?[0-9]+:[0-9]+:[ATCG]+:[ATCG]+", region_ref_to_check,  re.IGNORECASE)    
+                if matched_snpid is None:
+                    pass
+                else:
+                    lead_snpid = matched_snpid.group(0).split(":")
+                    if len(lead_snpid)==4:
+                        lead_chr= int(lead_snpid[0])
+                        lead_pos= int(lead_snpid[1])
+                        lead_ea= lead_snpid[2]
+                        lead_nea= lead_snpid[3]
+                        chrpos_match = (sumstats["CHR"] == lead_chr) & (sumstats["POS"] == lead_pos)
+                        eanea_match = ((sumstats["EA"] == lead_ea) & (sumstats["NEA"] == lead_nea)) | ((sumstats["EA"] == lead_nea) & (sumstats["NEA"] == lead_ea)) 
+                        if "rsID" in sumstats.columns:
+                            lead_id = sumstats.index[chrpos_match&eanea_match].to_list()
+                        if "SNPID" in sumstats.columns:
+                            lead_id = sumstats.index[chrpos_match&eanea_match].to_list()     
+                if type(lead_id) is list:
+                    if len(lead_id)>0:
+                        lead_id = int(lead_id[0])   
+                        log.warning("Trying matching variant {} using CHR:POS:EA:NEA to {}... ".format(region_ref_to_check,lead_id))
+
+        if type(lead_id) is list:
+            if len(lead_id)==0 :
+                log.warning("Extracting variant: {} not found in sumstats.. Skipping..".format(region_ref_to_check))
                 #lead_id = sumstats["scaled_P"].idxmax()
                 lead_id = None
                 return lead_id
@@ -275,7 +300,7 @@ def _get_lead_id(sumstats=None, region_ref=None, log=None, verbose=True):
 
     return lead_id
 
-def _pinpoint_lead(sumstats,ax1,region_ref, lead_color, marker_size, log, verbose,region_marker_shapes):
+def _pinpoint_lead(sumstats,ax1,region_ref, lead_color, marker_size, log, verbose, region_marker_shapes):
     
     if region_ref is None:
         log.write(" -Extracting lead variant..." , verbose=verbose)
@@ -416,6 +441,11 @@ def _plot_gene_track(
     texts_to_adjust_left = []
     texts_to_adjust_middle = []
     texts_to_adjust_right = []
+
+    
+    sig_gene_names=[]
+    sig_gene_lefts=[]
+    sig_gene_rights=[]
     for index,row in uniq_gene_region.iterrows():
 
         gene_color="#020080"
@@ -426,21 +456,18 @@ def _plot_gene_track(
             gene_anno = "<-" + row["name"] 
         
 
-        sig_gene_names=[]
-        sig_gene_lefts=[]
-        sig_gene_rights=[]
+
         for lead_snp_i in lead_snp_is:
             if region_lead_grid is True and lead_snp_i > gene_track_start_i+row["start"] and lead_snp_i < gene_track_start_i+row["end"] :
-                    gene_color=region_lead_grid_line["color"]
-                    sig_gene_names.append(row["name"])
-                    sig_gene_lefts.append(gene_track_start_i+row["start"])
-                    sig_gene_rights.append(gene_track_start_i+row["end"])
+                gene_color=region_lead_grid_line["color"]
+                sig_gene_names.append(row["name"])
+                sig_gene_lefts.append(gene_track_start_i+row["start"])
+                sig_gene_rights.append(gene_track_start_i+row["end"])
 
             # plot gene line
             ax3.plot((gene_track_start_i+row["start"],gene_track_start_i+row["end"]),
                         (row["stack"]*2,row["stack"]*2),color=gene_color,linewidth=linewidth_in_points/10)
         
-        
         # plot gene name
         if row["end"] >= region[2]:
             #right side
@@ -459,6 +486,7 @@ def _plot_gene_track(
     for index,row in exons.iterrows():
         exon_color="#020080" 
         for sig_gene_name, sig_gene_left, sig_gene_right in zip(sig_gene_names,sig_gene_lefts,sig_gene_rights):
+            
             if not pd.isnull(row["name"]):
                 if (region_lead_grid is True) and row["name"]==sig_gene_name:
                     exon_color = region_lead_grid_line["color"]  
@@ -468,7 +496,7 @@ def _plot_gene_track(
                 exon_color = region_lead_grid_line["color"]  
             else:
                 exon_color="#020080"
-
+        
         ax3.plot((gene_track_start_i+row["start"],gene_track_start_i+row["end"]),
                     (row["stack"]*2,row["stack"]*2),linewidth=linewidth_in_points*taf[3],color=exon_color,solid_capstyle="butt")
 
@@ -550,24 +578,42 @@ def process_vcf(sumstats,
             # figure out lead variant
             lead_id = _get_lead_id(sumstats, region_ref_single, log, verbose)
         
-        if lead_id is None:
-            sumstats[rsq] = None
-            sumstats[rsq] = sumstats[rsq].astype("float")
-            sumstats[ld_single] = 0             
-            continue
 
-        lead_pos = sumstats.loc[lead_id,pos]
+        lead_series = None
+        if lead_id is None:
+            
+            matched_snpid = re.match("(chr)?[0-9]+:[0-9]+:[ATCG]+:[ATCG]+",region_ref_single,  re.IGNORECASE)
+            
+            if matched_snpid is None:
+                sumstats[rsq] = None
+                sumstats[rsq] = sumstats[rsq].astype("float")
+                sumstats[ld_single] = 0    
+                continue    
+            else:
+                
+                lead_snpid = matched_snpid.group(0).split(":")[1:]
+                lead_pos = int(lead_snpid[0])
+                lead_snpid[0]= int(lead_snpid[0])
+                lead_series = pd.Series(lead_snpid)
+        else:
+            lead_pos = sumstats.loc[lead_id,pos]
 
+        
         # if lead pos is available: 
         if lead_pos in ref_genotype["variants/POS"]:
             
             # get ref index for lead snp
-            lead_snp_ref_index = match_varaint(sumstats.loc[lead_id,[pos,nea,ea]])
-            #lead_snp_ref_index = np.where(ref_genotype["variants/POS"] == lead_pos)[0][0]
+            if lead_series is None:
+                lead_snp_ref_index = match_varaint(sumstats.loc[lead_id,[pos,nea,ea]])
+                #lead_snp_ref_index = np.where(ref_genotype["variants/POS"] == lead_pos)[0][0]
+            else:
+                log.warning("Computing LD: {} not found in sumstats but found in reference...Still Computing...".format(region_ref_single))
+                lead_snp_ref_index = match_varaint(lead_series)
 
             # non-na other snp index
             other_snps_ref_index = sumstats["REFINDEX"].dropna().astype("int").values
             # get genotype 
+            
             lead_snp_genotype = GenotypeArray([ref_genotype["calldata/GT"][lead_snp_ref_index]]).to_n_alt()
             try:
                 if len(set(lead_snp_genotype[0]))==1:
@@ -604,10 +650,10 @@ def process_vcf(sumstats,
                 sumstats.loc[to_change_color,ld_single] = 1
             to_change_color = sumstats[rsq]>ld_threshold
             sumstats.loc[to_change_color,ld_single] = index+2
-
-        sumstats.loc[lead_id,ld_single] = len(region_ld_threshold)+2
         
-        sumstats.loc[lead_id,lead] = 1
+        if lead_series is None:
+            sumstats.loc[lead_id,ld_single] = len(region_ld_threshold)+2
+            sumstats.loc[lead_id,lead] = 1
 
     ####################################################################################################    
     final_shape_col = "SHAPE"