gwaslab 3.4.37__py3-none-any.whl → 3.4.39__py3-none-any.whl

gwaslab/g_Sumstats_summary.py

@@ -15,7 +15,7 @@ def summarize(insumstats,
     for i in [snpid,rsid,eaf,p,n,status]:
         if i in insumstats.columns:
             cols.append(i)
-    sumstats= insumstats.loc[:,cols].copy()
+    sumstats= insumstats[cols].copy()
     ###############################################################################
     numeric_cols=[]
     output = {}
@@ -68,7 +68,7 @@ def summarize(insumstats,
         sumstats.drop(columns='uniq_index',inplace=True)
         status_dic = {}
         for index,row in status_summary.iterrows():
-            status_dic[str(index)]=row[0]
+            status_dic[str(index)]=row.iloc[0]
         output["STATUS"]=status_dic
         numeric_cols.append("STATUS")
     df = pd.DataFrame.from_dict({(i,j): output[i][j] 
@@ -84,7 +84,7 @@ def summarize(insumstats,
     return df 
 
 def sum_status(id_to_use, sumstats):
-        results = sumstats.groupby("STATUS").count()
+        results = sumstats.groupby("STATUS",observed=True).count()
         results = results.loc[results[id_to_use]>0,:].sort_values(id_to_use,ascending=False)
         return results
     

gwaslab/g_version.py

@@ -3,10 +3,10 @@ import subprocess
 import os
 import numpy as np 
 
-def _show_version(log=Log()):
+def _show_version(log=Log(), verbose=True):
     # show version when loading sumstats
-    log.write("GWASLab v{} https://cloufield.github.io/gwaslab/".format(gwaslab_info()["version"]))
-    log.write("(C) 2022-2024, Yunye He, Kamatani Lab, MIT License, gwaslab@gmail.com")
+    log.write("GWASLab v{} https://cloufield.github.io/gwaslab/".format(gwaslab_info()["version"]),verbose=verbose)
+    log.write("(C) 2022-2024, Yunye He, Kamatani Lab, MIT License, gwaslab@gmail.com",verbose=verbose)
 
 def _get_version():
     # return short version string like v3.4.33
@@ -15,12 +15,12 @@ def _get_version():
 def gwaslab_info():
     # version meta information
     dic={
-   "version":"3.4.37",
-   "release_date":"20240129"
+   "version":"3.4.39",
+   "release_date":"20240210"
     }
     return dic   
 
-def _checking_plink_version(v=2,log=Log()):
+def _checking_plink_version(v=2,log=Log(), verbose=True):
     if v==1:
         which_plink_script = "plink --version"
     elif v==2:
@@ -29,19 +29,19 @@ def _checking_plink_version(v=2,log=Log()):
     log.write("   -PLINK version: {}".format(output.strip()))
     return log
 
-def _checking_r_version(r, log):
+def _checking_r_version(r, log=Log(), verbose=True):
     which_r_script = "{} --version".format(r)
     output = subprocess.check_output(which_r_script, stderr=subprocess.STDOUT, shell=True,text=True)
-    log.write(" -R version: {}".format(output.strip()))
+    log.write(" -R version: {}".format(output.strip()),verbose=verbose)
     return log
 
-def _check_susie_version(r,log):
+def _check_susie_version(r,log=Log(), verbose=True):
     rscript = 'print(packageVersion("susieR"))'
     temp_r = "_gwaslab_susie_temp_check_version_{}.R".format(np.random.randint(1, 99999999))
     with open(temp_r,"w") as file:
         file.write(rscript)
     which_susie_script = "{} {}".format(r, temp_r)
     output = subprocess.check_output(which_susie_script, stderr=subprocess.STDOUT, shell=True,text=True)
-    log.write(" -SuSieR version: {}".format(output.strip()))
+    log.write(" -SuSieR version: {}".format(output.strip()),verbose=verbose)
     os.remove(temp_r)
     return log

gwaslab/hm_casting.py

@@ -5,18 +5,24 @@ from pandas.api.types import CategoricalDtype
 from gwaslab.g_vchange_status import copy_status
 from gwaslab.g_vchange_status import vchange_status
 from gwaslab.qc_fix_sumstats import flipallelestats
+from gwaslab.qc_check_datatype import check_datatype
+from gwaslab.qc_fix_sumstats import start_to
 from gwaslab.util_in_fill_data import filldata
 from Bio import SeqIO
 from itertools import combinations
 
-def _merge_mold_with_sumstats(mold, sumstats, ref_path=None, windowsizeb=10, log=Log(),suffixes=("_MOLD",""),verbose=True,return_not_matched_mold =False):
+def _merge_mold_with_sumstats_by_chrpos(mold, sumstats, ref_path=None, windowsizeb=10, log=Log(),suffixes=("_MOLD",""),verbose=True,return_not_matched_mold =False):
+    
+
     cols_to_drop = []
     for i in sumstats.columns:
         if i in ["SNPID","rsID"]:
             cols_to_drop.append(i)
+
+    log.write("Start to merge sumstats...", verbose=verbose)
     
     if len(cols_to_drop)>0:
-        log.write("Dropping old IDs:{}".format(cols_to_drop))
+        log.write(" -Dropping old IDs:{}".format(cols_to_drop), verbose=verbose)
         sumstats = sumstats.drop(columns=cols_to_drop)
     
     if ref_path is not None :
@@ -29,18 +35,20 @@ def _merge_mold_with_sumstats(mold, sumstats, ref_path=None, windowsizeb=10, log
     if return_not_matched_mold:
         mold["_IDENTIFIER_FOR_VARIANT"] = range(len(mold))
 
+    # mold sumffix + mold 
     mold_sumstats = pd.merge(mold, sumstats, on=["CHR","POS"], how="inner",suffixes=suffixes)
-    log.write("After merging by CHR and POS:{}".format(len(mold_sumstats)))
+    log.write(" -After merging by CHR and POS:{}".format(len(mold_sumstats)), verbose=verbose)
     
     mold_sumstats = _keep_variants_with_same_allele_set(mold_sumstats,suffixes=suffixes)
-    log.write("Matched variants:{}".format(len(mold_sumstats)))
+
+    log.write(" -Matched variants:{}".format(len(mold_sumstats)), verbose=verbose)
     
-    if ref_path is not None:
-        # match removed sumstats
-        mold_removed = mold.loc[~mold[index1].isin(mold_sumstats[index1]),:]
-        iron_removed = sumstats.loc[~sumstats[index2].isin(mold_sumstats[index2]),:]
-        _match_two_sumstats(mold_removed,iron_removed,ref_path,windowsizeb=windowsizeb)
-        mold_sumstats.drop(columns=["_INDEX",""])
+    #if ref_path is not None:
+    #    # match removed sumstats
+    #    mold_removed = mold.loc[~mold[index1].isin(mold_sumstats[index1]),:]
+    #    iron_removed = sumstats.loc[~sumstats[index2].isin(mold_sumstats[index2]),:]
+    #    _match_two_sumstats(mold_removed,iron_removed,ref_path,windowsizeb=windowsizeb)
+    #    mold_sumstats.drop(columns=["_INDEX",""])
     
     if return_not_matched_mold == True:
         sumstats1 = mold.loc[~mold["_IDENTIFIER_FOR_VARIANT"].isin(mold_sumstats["_IDENTIFIER_FOR_VARIANT"]),:]
@@ -59,14 +67,17 @@ def _keep_variants_with_same_allele_set(sumstats, log=Log(),verbose=True,suffixe
 
     all_alleles = set(list(sumstats[ea1].unique())+list(sumstats[nea1].unique())+list(sumstats[ea2].unique())+list(sumstats[nea2].unique()))
     allele_type = CategoricalDtype(categories=all_alleles, ordered=False)
-    sumstats.loc[:, [nea1,ea1,nea2,ea2]] = sumstats.loc[:, [nea1,ea1,nea2,ea2]].astype(allele_type)
+    sumstats[[nea1,ea1,nea2,ea2]] = sumstats[[nea1,ea1,nea2,ea2]].astype(allele_type)
     
     is_perfect_match = (sumstats[ea2] == sumstats[ea1]) & (sumstats[nea2] == sumstats[nea1])
     is_flipped_match = (sumstats[ea2] == sumstats[nea1]) & (sumstats[nea2] == sumstats[ea1])
     is_allele_set_match = is_flipped_match | is_perfect_match
     
-    sumstats.loc[~is_allele_set_match,:]
-
+    log.write(" -Matching alleles and keeping only variants with same allele set: ", verbose=verbose)
+    log.write("  -Perfect match: {}".format(sum(is_perfect_match)), verbose=verbose)
+    log.write("  -Flipped match: {}".format(sum(is_flipped_match)), verbose=verbose)
+    log.write("  -Unmatched : {}".format(sum(~is_allele_set_match)), verbose=verbose)
+    
     return sumstats.loc[is_allele_set_match,:]
 
 def _align_with_mold(sumstats, log=Log(),verbose=True, suffixes=("_MOLD","")):
@@ -77,10 +88,18 @@ def _align_with_mold(sumstats, log=Log(),verbose=True, suffixes=("_MOLD","")):
     nea2="NEA"+suffixes[1]
     status1="STATUS"+suffixes[0]
     status2="STATUS"+suffixes[1]
+
     is_perfect_match = (sumstats[ea2] == sumstats[ea1]) & (sumstats[nea2] == sumstats[nea1])
     is_flipped_match = (sumstats[ea2] == sumstats[nea1]) & (sumstats[nea2] == sumstats[ea1])
     
+    log.write(" -Aligning alleles with reference: ", verbose=verbose)
+    log.write("  -Perfect match: {}".format(sum(is_perfect_match)), verbose=verbose)
+    log.write("  -Flipped match: {}".format(sum(is_flipped_match)), verbose=verbose)
+    
+    log.write("  -For perfect match: copy STATUS from reference...", verbose=verbose)
     sumstats.loc[is_perfect_match,status2] = copy_status(sumstats.loc[is_perfect_match,status1], sumstats.loc[is_perfect_match,status2],6)
+    
+    log.write("  -For Flipped match: convert STATUS xxxxx[456789]x to xxxxx3x...", verbose=verbose)
     sumstats.loc[is_flipped_match,status2] = vchange_status(sumstats.loc[is_flipped_match,status2],6,"456789","333333")
     
     return sumstats
@@ -119,9 +138,9 @@ def _sort_pair_cols(molded_sumstats, verbose=True, log=Log(), order=None, stats_
         if i not in order: 
             output_columns.append(i)
     
-    if verbose: log.write(" -Reordering columns to    :", ",".join(output_columns))
-    molded_sumstats = molded_sumstats.loc[:, output_columns]
-    if verbose: log.write("Finished sorting columns successfully!")
+    log.write(" -Reordering columns to    :", ",".join(output_columns), verbose=verbose)
+    molded_sumstats = molded_sumstats[ output_columns]
+    log.write("Finished sorting columns successfully!", verbose=verbose)
     
     return molded_sumstats
 
@@ -154,7 +173,7 @@ def _match_two_sumstats(mold,sumstats,ref_path,windowsizeb=25,verbose=True,log=L
             record_chr = int(str(record.id).strip("chrCHR").upper())
             
             if record_chr in chromlist:
-                if verbose:  log.write(record_chr," ", end="",show_time=False) 
+                log.write(record_chr," ", end="",show_time=False,verbose=verbose) 
                 chromlist.remove(record_chr)
             else:
                 continue