gwaslab 3.5.7__py3-none-any.whl → 3.6.0__py3-none-any.whl

gwaslab/util_ex_run_ccgwas.py

@@ -0,0 +1,155 @@
+import subprocess
+import os
+import gc
+import pandas as pd
+import numpy as np
+from gwaslab.g_Log import Log
+from gwaslab.g_version import _checking_r_version
+from gwaslab.g_version import _check_susie_version
+from gwaslab.qc_fix_sumstats import start_to
+from gwaslab.qc_fix_sumstats import finished
+from gwaslab.util_ex_calculate_ldmatrix import _extract_variants_in_locus
+from gwaslab.util_in_get_sig import getsig
+
+def _run_ccgwas(    sumstats_pair, 
+                    r="Rscript",
+                    group="Group1",
+                    studies=None,
+                    traits=None,
+                    meta=None,
+                    ldsc=None,
+                    ldsc_rg=None,
+                    nstudy=2,
+                    K_A1A0 = None , 
+                    K_A1A0_high = None , 
+                    K_A1A0_low = None ,  
+                    K_B1B0 = None , 
+                    K_B1B0_high = None , 
+                    K_B1B0_low = None , 
+                    h2l_A1A0 = None , 
+                    h2l_B1B0 = None , 
+                    rg_A1A0_B1B0 = None , 
+                    intercept_A1A0_B1B0 = None , 
+                    m = 1e4,
+                    N_A1 = None,
+                    N_B1 = None,
+                    N_A0 = None , 
+                    N_B0 = None , 
+                    N_overlap_A0B0 = 0,
+                    log=Log(), 
+                    verbose=True):
+    
+    log.write(" Start to run CCGWAS from command line:", verbose=verbose)
+    log.write(" -Methods: : {}...".format("Peyrot, W. J., & Price, A. L. (2021). Identifying loci with different allele frequencies among cases of eight psychiatric disorders using CC-GWAS. Nature genetics, 53(4), 445-454."),verbose=verbose)
+    #"SNP, CHR, BP, EA, NEA, FRQ, OR, SE, P, Neff"
+    log.write(" -Running CCGWAS for: {}...".format(group),verbose=verbose)
+    
+    snp_info_cols=["SNPID","CHR","POS","EA","NEA"]
+    stats_cols=["EAF","OR","SE","P","N_EFF"]
+
+    if meta["gwaslab"]["objects"][0]["gwaslab"]["population_prevalence"] != "Unknown":
+          K_A1A0 = float(meta["gwaslab"]["objects"][0]["gwaslab"]["population_prevalence"])
+          K_A1A0_high = 1.1 * K_A1A0
+          K_A1A0_low = K_A1A0/ 1.1
+    
+    if meta["gwaslab"]["objects"][1]["gwaslab"]["population_prevalence"] != "Unknown":
+          K_B1B0 = float(meta["gwaslab"]["objects"][1]["gwaslab"]["population_prevalence"])
+          K_B1B0_high = 1.1 * K_B1B0
+          K_B1B0_low = K_B1B0/ 1.1
+
+    if h2l_A1A0 is None: 
+        h2l_A1A0 = ldsc[0].loc[0, "h2_liab"]
+    if h2l_B1B0 is None: 
+        h2l_B1B0 = ldsc[1].loc[0, "h2_liab"]
+    if rg_A1A0_B1B0 is None:
+        rg_A1A0_B1B0 = ldsc_rg.loc[(ldsc_rg["p1"]==studies[0])&(ldsc_rg["p2"]==studies[1]), :].iloc[0,ldsc_rg.columns.get_loc("rg")]
+    if intercept_A1A0_B1B0 is None:
+        intercept_A1A0_B1B0 = ldsc_rg.loc[(ldsc_rg["p1"]==studies[0])&(ldsc_rg["p2"]==studies[1]), :].iloc[0,ldsc_rg.columns.get_loc("gcov_int")]
+
+    # prepare input files sumstats_multi
+    for i in range(nstudy):
+        output_cols = snp_info_cols + list(map(lambda x: x+"_{}".format(i+1), stats_cols))
+
+        dic= {"SNPID":"SNP",
+              "POS":"BP",
+              "EAF_{}".format(i+1):"FRQ",
+              "OR_{}".format(i+1):"OR",
+              "SE_{}".format(i+1):"SE",
+              "P_{}".format(i+1):"P",
+              "N_EFF_{}".format(i+1):"Neff"
+              }
+
+        sumstats_pair[output_cols].rename(columns=dic).to_csv("./{}_{}.txt.gz".format(group, studies[i]),index=None,sep="\t")
+    
+    output_prefix = "{group}_ccgwas".format(group=group)
+    r_log=""
+    log = _checking_r_version(r, log)
+
+    rscript='''
+library(data.table)
+library(R.utils)
+library(CCGWAS)
+
+CCGWAS( outcome_file = "{output_prefix}" , 
+        A_name = "{study1}" , 
+        B_name = "{study2}" , 
+        sumstats_fileA1A0 = "./{group}_{study1}.txt.gz" ,
+        sumstats_fileB1B0 = "./{group}_{study2}.txt.gz" ,
+        K_A1A0 = {K_A1A0} , 
+        K_A1A0_high = {K_A1A0_high} , 
+        K_A1A0_low = {K_A1A0_low},  
+
+        K_B1B0 ={K_B1B0} , 
+        K_B1B0_high ={K_B1B0_high} , 
+        K_B1B0_low = {K_B1B0_low} , 
+
+        h2l_A1A0 ={h2l_A1A0}, 
+        h2l_B1B0 = {h2l_B1B0} , 
+        rg_A1A0_B1B0 = {rg_A1A0_B1B0} , 
+
+        intercept_A1A0_B1B0 = {intercept_A1A0_B1B0} , 
+        m = {m} ,  
+        N_A1 =  {N_A1} , 
+        N_B1 =  {N_B1} , 
+        N_A0 =  {N_A0} , 
+        N_B0 =  {N_B0} , 
+        N_overlap_A0B0 =  {N_overlap_A0B0} )
+        '''.format(
+            output_prefix=output_prefix,
+            study1=studies[0],
+            study2=studies[1],
+            group=group,
+            K_A1A0 = K_A1A0 , 
+            K_A1A0_high = K_A1A0_high , 
+            K_A1A0_low = K_A1A0_low ,  
+            K_B1B0 =K_B1B0 , 
+            K_B1B0_high = K_B1B0_high , 
+            K_B1B0_low =K_B1B0_low , 
+            h2l_A1A0 = h2l_A1A0 , 
+            h2l_B1B0 = h2l_B1B0 , 
+            rg_A1A0_B1B0 = rg_A1A0_B1B0 , 
+            intercept_A1A0_B1B0 = intercept_A1A0_B1B0 , 
+            m = m,
+            N_A1 = N_A1,
+            N_B1 = N_B1,
+            N_A0 = N_A0 , 
+            N_B0 = N_B0 , 
+            N_overlap_A0B0 = N_overlap_A0B0,
+        )
+
+    with open("_{}_gwaslab_ccgwas_temp.R".format( group),"w") as file:
+            file.write(rscript)
+
+    script_run_r = "{} _{}_gwaslab_ccgwas_temp.R".format(r, group )
+    
+    try:
+        log.write(" -Running CCGWAS from command line...", verbose=verbose)
+        output = subprocess.check_output(script_run_r, stderr=subprocess.STDOUT, shell=True,text=True)
+        r_log+= output + "\n"
+        #os.remove("_{}_{}_gwaslab_hyprcoloc_temp.R".format(study,locus))
+    except subprocess.CalledProcessError as e:
+        log.write(e.output)
+        #os.remove("_{}_{}_gwaslab_hyprcoloc_temp.R".format(study,locus))
+    log.write(" -Finishing CCGWAS for {}...".format(group),verbose=verbose)
+    log.write("Finished Case-case GWAS using CCGWAS.", verbose=verbose)
+    return output_prefix

gwaslab/util_ex_run_coloc.py

@@ -58,7 +58,7 @@ def _run_coloc_susie(filepath, r="Rscript",
         rscript='''
         library(coloc)
         
-        df = read.csv("{sumstats_path}",header=TRUE)
+        df = read.csv("{sumstats_path}",sep="\t",header=TRUE)
 
         R <- as.matrix(read.csv("{ld_r_matrix_path}",sep="\t",header=FALSE))
 

gwaslab/util_ex_run_hyprcoloc.py

@@ -0,0 +1,117 @@
+import subprocess
+import os
+import gc
+import pandas as pd
+import numpy as np
+from gwaslab.g_Log import Log
+from gwaslab.g_version import _checking_r_version
+from gwaslab.g_version import _check_susie_version
+from gwaslab.qc_fix_sumstats import start_to
+from gwaslab.qc_fix_sumstats import finished
+from gwaslab.util_ex_calculate_ldmatrix import _extract_variants_in_locus
+from gwaslab.util_in_get_sig import getsig
+
+def _run_hyprcoloc(  sumstats_multi, 
+                     r="Rscript",
+                     study="Group1",
+                     traits=None,
+                     types=None, 
+                     loci=None,
+                     nstudy=2,
+                     windowsizekb=1000,
+                     build="99",
+                     log=Log(), 
+                     verbose=True):
+    
+    log.write(" Start to run hyprcoloc from command line:", verbose=verbose)
+    if traits is None:
+         traits_to_form_string = [ '"trait_{}"'.format(i+1) for i in range(nstudy)]
+    else:
+         traits_to_form_string = ['"{}"'.format(i) for i in traits]
+    
+    hyprcoloc_res_combined = pd.DataFrame()
+
+    if loci is None:
+        log.write(" -Loci were not provided. All significant loci will be automatically extracted...",verbose=verbose)
+        sig_df = getsig(sumstats_multi,id="SNPID",chrom="CHR",pos="POS",p="P_MIN", build=build)
+    else:
+        sig_df = sumstats_multi.loc[sumstats_multi["SNPID"].isin(loci),:]
+
+    for index,row  in  sig_df.iterrows():
+    #row = sig_df.iloc[0,:]
+        
+        # extract locus
+        locus = row["SNPID"]
+        log.write(" -Running hyprcoloc for locus : {}...".format(locus),verbose=verbose)
+
+        # prepare input files sumstats_multi
+        output_beta_cols = []
+        output_se_cols = []
+
+        for i in range(nstudy):
+            output_beta_cols.append("BETA_{}".format(i+1))
+            output_se_cols.append("SE_{}".format(i+1))
+
+        matched_sumstats = _extract_variants_in_locus(sumstats_multi, windowsizekb, locus = (row["CHR"],row["POS"]))
+        
+        to_export = matched_sumstats[["SNPID"] + output_se_cols + output_beta_cols].dropna()
+        
+        if len(to_export)>0:
+            log.write(" -Number of shared variants in locus {} : {}...".format(locus, len(to_export)),verbose=verbose)
+            to_export[["SNPID"] + output_beta_cols].to_csv("{}_{}_beta_cols.tsv.gz".format(study,locus), index=None,sep="\t")
+            to_export[["SNPID"] + output_se_cols].to_csv("{}_{}_se_cols.tsv.gz".format(study,locus), index=None,sep="\t")
+        else:
+            log.write(" -No shared variants in locus {}...skipping".format(locus),verbose=verbose)
+            continue
+
+        r_log=""
+        log = _checking_r_version(r, log)
+
+        rscript='''
+library(hyprcoloc)
+
+betas<-read.csv("{study}_{locus}_beta_cols.tsv.gz",row.names = 1,sep="\\t")
+ses  <-read.csv("{study}_{locus}_se_cols.tsv.gz",row.names = 1,sep="\\t")
+
+betas <- as.matrix(betas)
+ses <- as.matrix(ses)
+
+traits <- c({traits_string})
+snpid <- rownames(betas)
+
+res <- hyprcoloc(betas, 
+          ses, 
+          trait.names = traits, 
+          snp.id = snpid,
+          snpscore=TRUE)
+
+write.csv(res[[1]], "{study}_{locus}_{nstudy}studies.res",row.names = FALSE) 
+        '''.format(
+            study=study,
+            locus=locus,
+            nstudy=nstudy,
+            traits_string = ','.join(traits_to_form_string)
+        )
+
+        output_path = "{study}_{locus}_{nstudy}studies.res".format(study=study, locus=locus, nstudy=nstudy)
+        output_prefix = "{study}_{locus}_{nstudy}studies".format(study=study, locus=locus, nstudy=nstudy)
+        
+        with open("_{}_{}_gwaslab_hyprcoloc_temp.R".format(study,locus),"w") as file:
+                file.write(rscript)
+
+        script_run_r = "{} _{}_{}_gwaslab_hyprcoloc_temp.R".format(r, study,locus)
+        
+        try:
+            log.write(" Running hyprcoloc from command line...", verbose=verbose)
+            output = subprocess.check_output(script_run_r, stderr=subprocess.STDOUT, shell=True,text=True)
+            r_log+= output + "\n"
+            #os.remove("_{}_{}_gwaslab_hyprcoloc_temp.R".format(study,locus))
+            hyprcoloc_res = pd.read_csv(output_path)    
+            hyprcoloc_res["PREFIX"] = output_prefix 
+            hyprcoloc_res_combined = pd.concat([hyprcoloc_res_combined, hyprcoloc_res],ignore_index=True)
+        except subprocess.CalledProcessError as e:
+            log.write(e.output)
+            #os.remove("_{}_{}_gwaslab_hyprcoloc_temp.R".format(study,locus))
+        log.write(" -Finishing hyprcoloc for locus : {}...".format(locus),verbose=verbose)
+    log.write("Finished clocalization using hyprcoloc.", verbose=verbose)
+    return hyprcoloc_res_combined

gwaslab/util_ex_run_magma.py

@@ -0,0 +1,74 @@
+import subprocess
+import os
+import gc
+import pandas as pd
+import numpy as np
+from gwaslab.g_Log import Log
+from gwaslab.util_in_filter_value import _exclude_hla
+
+def _run_magma(sumstats, 
+                magma="magma",
+                study="Study1",
+                exclude_hla=True,
+                window="35,10",
+                id_to_use="rsID",
+                ref=None,
+                ncbi=None,
+                set_annot=None,
+                out="./",
+                delete=True, 
+                ncol="N",
+                build="19",
+                log=Log(), 
+                verbose=True):
+    
+    log.write(" Start to run magma from command line:", verbose=verbose)
+
+    if exclude_hla==True:
+        sumstats = _exclude_hla(sumstats, build =build)
+
+    snploc="{}{}.rsid.chr.pos.tsv".format(out,study)
+    pval="{}{}.rsid.p.n.tsv".format(out, study)
+
+    log.write(f" -writing temp file for --snp-loc:{snploc}", verbose=verbose)
+    sumstats.dropna()[[id_to_use,"CHR","POS"]].rename(columns={id_to_use:"SNP"}).to_csv("{}{}.rsid.chr.pos.tsv".format(out,study),index=None, sep="\t")
+    
+    log.write(f" -writing temp file for --pval:{pval}", verbose=verbose)
+    sumstats.dropna()[[id_to_use,"P","N"]].rename(columns={id_to_use:"SNP"}).to_csv("{}{}.rsid.p.n.tsv".format(out,study),index=None, sep="\t")
+    
+    log.write(f" --annotate window: {window}", verbose=verbose)
+    log.write(f" --gene-loc: {ncbi}", verbose=verbose)
+    log.write(f" --bfile: {ref}", verbose=verbose)
+    log.write(f" Output prefix: {out}", verbose=verbose)
+
+    bash_script=f'''
+
+#!/bin/bash
+
+{magma} --annotate window={window}  --snp-loc {snploc} --gene-loc {ncbi} --out {study}
+
+{magma} --bfile {ref} --pval {pval} ncol={ncol} --gene-annot {study}.genes.annot --out {study}
+
+'''
+    
+    if set_annot is not None:
+        bash_script+=f'''
+{magma} --gene-results {study}.genes.raw --set-annot {set_annot} --out {study}
+'''
+    log.write(f"Script: {bash_script}")
+
+    try:
+        log.write(" Running magma from command line...", verbose=verbose)
+        output = subprocess.check_output(bash_script, stderr=subprocess.STDOUT, shell=True,text=True)
+        output =  output + "\n"
+        
+        if delete == True:
+            os.remove(snploc)
+            os.remove(pval)
+
+    except subprocess.CalledProcessError as e:
+        log.warning("ERROR!")
+        log.write(e.output)
+
+    log.write("Finished running magma.", verbose=verbose)
+

gwaslab/util_ex_run_mesusie.py

@@ -0,0 +1,155 @@
+import subprocess
+import os
+import gc
+import pandas as pd
+import numpy as np
+from gwaslab.g_Log import Log
+from gwaslab.g_version import _checking_r_version
+from gwaslab.g_version import _check_susie_version
+from gwaslab.qc_fix_sumstats import start_to
+from gwaslab.qc_fix_sumstats import finished
+from gwaslab.viz_plot_stackedregional import _sort_args
+
+def _run_mesusie(filepath, 
+                 r="Rscript",
+                 types=None, ns=None, 
+                 fillldna=True, delete=False, 
+                 coloc_args="", 
+                 susie_args="", 
+                 ncols=None,
+                 d1_args="",
+                 d2_args="",
+                 log=Log(), 
+                 verbose=True):
+    
+    log.write(" Start to run mesusie from command line:", verbose=verbose)
+    pass
+
+    if ns is None:
+        if ncols is not None:
+            ns = ncols
+    log.write(" -Ns: {} and {}".format(ns[0],ns[1]), verbose=verbose)
+
+    if filepath is None:
+        log.write(" -File path is None.", verbose=verbose)
+        log.write("Finished finemapping using MESuSie.", verbose=verbose)
+        return pd.DataFrame()
+        
+    filelist = pd.read_csv(filepath,sep="\t")
+    r_log=""
+    # write R script
+    locus_pip_cs = pd.DataFrame()
+
+    log = _checking_r_version(r, log)
+    #log = _check_susie_version(r,log)
+    r_script_init='''
+library(MESuSiE)
+ld_list <- list()
+summ_stat_list <- list()
+    '''
+    r_scripts_for_loading =[r_script_init]
+
+    
+    for index, row in filelist.iterrows(): 
+        gc.collect()
+        if index==0:
+            study0 = row["STUDY"]
+        study = row["STUDY"]
+        group = row["GROUP"]
+        ld_r_matrix = row["LD_R_MATRIX"]
+        sumstats = row["LOCUS_SUMSTATS"]
+        locus=row["LOCUS"]
+
+        log.write(" -Running for: {} - {}".format(row["SNPID"],row["STUDY"] ), verbose=verbose)
+        log.write("  -Locus sumstats:{}".format(sumstats), verbose=verbose)
+        log.write("  -LD r matrix:{}".format(ld_r_matrix), verbose=verbose)
+
+        rscript='''
+sum{index} <-  read.csv("{sumstats}",sep="\\t")
+sum{index}$Z <- sum{index}$Beta/sum{index}$Se
+sum{index}$N <- {n}
+ld{index} <- read.csv("{ld_r_matrix}",sep="\\t",header=FALSE)
+ld{index}[is.na(ld{index})]  <- 0
+names(ld{index}) <- sum{index}$SNP
+ld_list${study} <-  as.matrix(ld{index})
+summ_stat_list${study} <- sum{index}
+
+png(filename="./diagnostic_{group}_{locus}_{index}.png")
+diagnostic <- kriging_rss(summ_stat_list${study}$Z, ld_list${study})
+diagnostic$plot
+dev.off()
+        '''.format(
+             index = index,
+             study = study,
+             group=group,
+             locus = locus,
+             n = ns[index],
+             sumstats = sumstats,
+             ld_r_matrix = ld_r_matrix
+        )
+        r_scripts_for_loading.append(rscript)
+    
+    rscript_loading = "".join(r_scripts_for_loading)
+    
+
+    rscript_computing='''
+MESuSiE_res<-meSuSie_core(ld_list, summ_stat_list, L=10)'''
+
+    rscript_output = '''
+saveRDS(MESuSiE_res, file = "{group}_{locus}.rds")
+pips <- cbind(summ_stat_list${study0}$SNP, summ_stat_list${study0}$CHR, summ_stat_list${study0}$POS, MESuSiE_res$pip_config)
+colnames(pips)[1] <-"SNPID"
+colnames(pips)[2] <-"CHR"
+colnames(pips)[3] <-"POS"
+pips <- data.frame(pips)
+pips[c("CREDIBLE_SET_INDEX")] <- 0 
+pips[c("CS_CATEGORY")] <- NA
+for (i in 1:length(MESuSiE_res$cs$cs)) {{
+    pips[MESuSiE_res$cs$cs[[i]],c("CREDIBLE_SET_INDEX")]<-i
+    pips[MESuSiE_res$cs$cs[[i]],c("CS_CATEGORY")] <- MESuSiE_res$cs$cs_category[[i]]
+}}
+write.csv(pips, "{group}_{locus}.pipcs", row.names = FALSE)
+
+write.csv(MESuSiE_res$cs$cs_index, "{group}_{locus}.cscs_index", row.names = FALSE)
+write.csv(MESuSiE_res$cs$purity, "{group}_{locus}.cspurity", row.names = FALSE)
+write.csv(MESuSiE_res$cs$cs_category, "{group}_{locus}.cscs_category", row.names = FALSE)
+
+
+for (p in MESuSiE_res$cs$cs) {{
+  write(p,"{group}_{locus}.cscs_i", append=TRUE, sep="\t", ncolumns=10000000)
+  write(summ_stat_list${study0}$SNP[p],"{group}_{locus}.cscs_snpid", append=TRUE, sep="\t", ncolumns=10000000)
+}}
+
+    '''.format(group=group,locus=locus,study0=study0)
+    
+    rscript_plotting='''
+png(filename="./{group}_{locus}_stacked_regions.png")
+MESuSiE_Plot(MESuSiE_res, ld_list ,summ_stat_list)
+dev.off()
+'''.format(group=group,locus=locus)
+    
+    rscript = rscript_loading + rscript_computing + rscript_output + rscript_plotting
+
+    log.write("  -MESuSie script: {}".format(rscript_computing), verbose=verbose)
+    
+    with open("_{}_{}_gwaslab_mesusie_temp.R".format(group,row["SNPID"]),"w") as file:
+            file.write(rscript)
+
+    script_run_r = "{} _{}_{}_gwaslab_mesusie_temp.R".format(r, group,row["SNPID"])
+    
+    try:
+        output = subprocess.check_output(script_run_r, stderr=subprocess.STDOUT, shell=True,text=True)
+        #plink_process = subprocess.Popen("exec "+script_run_r, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True,text=True)
+        #output1,output2 = plink_process.communicate()
+        #output= output1 + output2+ "\n"
+        #plink_process.kill()
+        log.write(" Running MESuSie from command line...", verbose=verbose)
+        r_log+= output + "\n"
+        
+        #os.remove("_{}_{}_gwaslab_coloc_susie_temp.R".format(study,row["SNPID"]))
+        
+    except subprocess.CalledProcessError as e:
+        log.write(e.output)
+        #os.remove("_{}_{}_gwaslab_coloc_susie_temp.R".format(study,row["SNPID"]))
+    log.write("Finished cross ancestry finemapping using MESuSie.", verbose=verbose)
+    return "./{}_@.pipcs".format(group)

gwaslab/util_ex_run_mtag.py

@@ -0,0 +1,92 @@
+import subprocess
+import os
+import gc
+import pandas as pd
+import numpy as np
+from gwaslab.g_Log import Log
+
+def _run_mtag(       sumstats_multi, 
+                     python="Rscript",
+                     mtag="",
+                     study="Group1",
+                     traits=None,
+                     out_prefix=None,
+                     types=None, 
+                     n_min=0,
+                     loci=None,
+                     nstudy=2,
+                     windowsizekb=1000,
+                     build="99",
+                     log=Log(), 
+                     verbose=True):
+    
+    log.write("Start to run MTAG from command line:", verbose=verbose)
+
+    if traits is None:
+         traits_to_form_string = [ 'trait_{}'.format(i+1) for i in range(nstudy)]
+    else:
+         traits_to_form_string = ['{}'.format(i) for i in traits]
+    
+    res_combined = pd.DataFrame()
+    # snpid    chr    bpos    a1    a2    freq    z    pval    n
+    
+    output_snp_info_cols =["rsID","CHR","POS","EA","NEA"]
+    sumstats_paths = []
+    for i in range(nstudy):
+        output_stats_cols=[]
+        for col in ["Z","P","EAF","N"]:
+            output_stats_cols.append("{}_{}".format(col, i+1))
+
+        rename_dict = {
+             "rsID":"snpid",
+             "CHR":"chr",
+             "POS":"bpos",
+             "EA":"a1",
+             "NEA":"a2",
+             "EAF_{}".format( i+1) :"freq",
+             "Z_{}".format( i+1)   :"z",
+             "P_{}".format( i+1)   :"pval",
+             "N_{}".format( i+1)   :"n",
+
+        }
+    
+        sumstats_multi[output_snp_info_cols+ output_stats_cols].rename(columns=rename_dict).to_csv("{}_{}.tsv.gz".format(study, traits_to_form_string[i]), index=None,sep="\t")
+        sumstats_paths.append("{}_{}.tsv.gz".format(study, traits_to_form_string[i]))
+    
+    python_log=""
+    if out_prefix is None:
+        out_prefix = "./{study}_{nstudy}studies".format(study=study, nstudy=nstudy)
+    
+    script='''
+{python} {mtag} \
+--sumstats {sumstats_paths_string} \
+--out {out_prefix} \
+--n_min {n_min} \
+--stream_stdout &
+        '''.format(
+            python=python,
+            n_min=n_min,
+            mtag=mtag,
+            out_prefix=out_prefix,
+            sumstats_paths_string = ",".join(sumstats_paths)
+        )
+    log.write(" MTAG script: {} ".format(script), verbose=verbose)
+
+    
+    with open("_{}_gwaslab_mtag_temp.sh".format(study),"w") as file:
+            file.write(script)
+    
+    os.chmod("_{}_gwaslab_mtag_temp.sh".format(study), 0o700)   
+
+    script_run = "./_{}_gwaslab_mtag_temp.sh".format(study)
+    
+    try:
+        log.write(" Running MTAG from command line...", verbose=verbose)
+        output = subprocess.check_output(script_run, stderr=subprocess.STDOUT, shell=True,text=True)
+        log.write(output)
+        python_log+= output + "\n"
+
+    except subprocess.CalledProcessError as e:
+        log.write(e.output)
+
+    log.write("Finished MTAG.", verbose=verbose)

gwaslab/util_ex_run_prscs.py

@@ -0,0 +1,85 @@
+#!/usr/bin/env python
+
+"""
+PRS-CS: a polygenic prediction method that infers posterior SNP effect sizes under continuous shrinkage (CS) priors
+using GWAS summary statistics and an external LD reference panel.
+
+Reference: T Ge, CY Chen, Y Ni, YCA Feng, JW Smoller. Polygenic Prediction via Bayesian Regression and Continuous Shrinkage Priors.
+           Nature Communications, 10:1776, 2019.
+
+
+Usage:
+python PRScs.py --ref_dir=PATH_TO_REFERENCE --bim_prefix=VALIDATION_BIM_PREFIX --sst_file=SUM_STATS_FILE --n_gwas=GWAS_SAMPLE_SIZE --out_dir=OUTPUT_DIR
+                [--a=PARAM_A --b=PARAM_B --phi=PARAM_PHI --n_iter=MCMC_ITERATIONS --n_burnin=MCMC_BURNIN --thin=MCMC_THINNING_FACTOR
+                 --chrom=CHROM --write_psi=WRITE_PSI --write_pst=WRITE_POSTERIOR_SAMPLES --seed=SEED]
+
+"""
+
+
+import os
+import sys
+import getopt
+
+import gwaslab.prscs_parse_genet as parse_genet
+import gwaslab.prscs_mcmc_gtb as mcmc_gtb
+import gwaslab.prscs_gigrnd as gigrnd
+
+
+def _run_prscs(
+         ref_dir=None,
+         bim_prefix=None,
+         sst_file=None,
+            a= 1, 
+            b= 0.5, 
+            phi= None, 
+            n_gwas= None,
+            n_iter= 1000, 
+            n_burnin= 500, 
+            thin= 5, 
+            out_dir= "./", 
+            chrom= range(1,23),
+            beta_std= 'FALSE', 
+            write_psi= 'FALSE', 
+            write_pst= 'FALSE', 
+            seed= None,
+            log=None,
+         **kwargs):
+    ## T Ge, CY Chen, Y Ni, YCA Feng, JW Smoller. Polygenic Prediction via Bayesian Regression and Continuous Shrinkage Priors.Nature Communications, 10:1776, 2019.
+    sst_file = sst_file.rename(columns={"rsID":"SNP","POS":"BP"})
+    log.write("Start to runnig PRScs...")
+    param_dict = {'ref_dir': ref_dir, 
+                    'bim_prefix': bim_prefix, 
+                    'a': a, 
+                    'b': b, 
+                    'phi': phi, 
+                    'n_gwas': n_gwas,
+                    'n_iter': n_iter, 
+                    'n_burnin': n_burnin, 
+                    'thin':thin, 
+                    'out_dir': out_dir, 
+                    'chrom': chrom,
+                    'beta_std': beta_std, 
+                    'write_psi': write_psi, 
+                    'write_pst': write_pst, 
+                    'seed': seed}
+    
+    for chrom in param_dict['chrom']:
+        log.write('##### process chromosome %d #####' % int(chrom))
+
+        if '1kg' in os.path.basename(param_dict['ref_dir']):
+            ref_dict = parse_genet.parse_ref(param_dict['ref_dir'] + '/snpinfo_1kg_hm3', int(chrom), log)
+        elif 'ukbb' in os.path.basename(param_dict['ref_dir']):
+            ref_dict = parse_genet.parse_ref(param_dict['ref_dir'] + '/snpinfo_ukbb_hm3', int(chrom), log)
+
+        vld_dict = parse_genet.parse_bim(param_dict['bim_prefix'], int(chrom))
+
+        sst_dict = parse_genet.parse_sumstats(ref_dict, vld_dict, sst_file, param_dict['n_gwas'], log)
+
+        ld_blk, blk_size = parse_genet.parse_ldblk(param_dict['ref_dir'], sst_dict, int(chrom), log)
+
+        mcmc_gtb.mcmc(param_dict['a'], param_dict['b'], param_dict['phi'], sst_dict, param_dict['n_gwas'], ld_blk, blk_size,
+            param_dict['n_iter'], param_dict['n_burnin'], param_dict['thin'], int(chrom), param_dict['out_dir'], param_dict['beta_std'],
+	    param_dict['write_psi'], param_dict['write_pst'], param_dict['seed'], log)
+
+    log.write("Finished!")
+