Beispiel #1
0
def _robustness_test_(all_snps, phenVals, outputFile, filter=0.1):
    import analyzePhenotype
    import analyzeHaplotype
    import random

    new_all_snps = []
    for snp in all_snps:
        if snp.count(0) > 1 and snp.count(1) > 1:
            new_all_snps.append(snp)
    print "Filtered", len(all_snps) - len(
        new_all_snps), " with minor allele count <2."
    all_snps = new_all_snps

    def getLeaveOneOutK(K, i):
        l = range(0, len(K))
        l.pop(i)
        new_k = numpy.core.take(K, l, 0)
        new_k = numpy.core.take(new_k, l, 1)
        return new_k

    print "Calculating kinship"
    t1 = time.time()
    K = calcKinship(all_snps)
    t2 = time.time()
    print "Took", t2 - t1, "seconds."
    """
	Leave one out test..
	"""
    if filter < 1.0:
        snps = random.sample(all_snps, int(len(all_snps) * filter))
        print "Number of SNPs:", len(snps)
    else:
        snps = all_snps


#	K = calcKinship(snps)

    print "running EMMA"
    t1 = time.time()
    true_pvals = _runEmma_(snps, phenVals, K)["ps"]
    true_pvals = map(float, true_pvals)
    t2 = time.time()
    print "Took", t2 - t1, "seconds."

    log_true_pvals = []
    for pval in true_pvals:
        log_true_pvals.append(-math.log(pval, 10))

    perm_pvalues_list = []
    for i in range(0, len(phenVals)):
        newPhenvals = phenVals[:]
        newPhenvals.pop(i)

        newSNPs = []
        for snp in snps:
            newSNP = snp[:]
            newSNP.pop(i)
            newSNPs.append(newSNP)

        new_k = getLeaveOneOutK(K, i)

        print "running EMMA"
        t1 = time.time()
        pvals = _runEmma_(newSNPs, newPhenvals, new_k)["ps"]
        pvals = map(float, pvals)
        t2 = time.time()
        print "Took", t2 - t1, "seconds."

        perm_pvalues_list.append(pvals)

    delta_pvals_list = []
    delta_log_pvals_list = []
    for perm_pvals in perm_pvalues_list:
        log_pvals = []
        delta_pvals = []
        delta_log_pvals = []
        for i in range(0, len(true_pvals)):
            pval = perm_pvals[i]
            true_pval = true_pvals[i]
            delta_pvals.append(true_pval - pval)

            log_true_pval = log_true_pvals[i]
            log_pval = -math.log(pval, 10)
            log_pvals.append(log_pval)
            delta_log_pvals.append(log_true_pval - log_pval)

        delta_pvals_list.append(delta_pvals)
        delta_log_pvals_list.append(delta_log_pvals)

    sd_log_pvals = []
    sd_pvals = []
    t_delta_log_pvals_list = map(list, zip(*delta_log_pvals_list))
    t_delta_pvals_list = map(list, zip(*delta_pvals_list))
    for i in range(0, len(true_pvals)):
        sd_log_pvals.append(util.calcSD(t_delta_log_pvals_list[i]))
        sd_pvals.append(util.calcSD(t_delta_pvals_list[i]))

    #Write SDs out to file, to be able to replot, or plot together with other methods... etc
    import csv
    sd_log_pval_file = outputFile + ".rob.log_pvals_sd"
    f = open(sd_log_pval_file, "w")
    w = csv.writer(f)
    w.writerow(["log_true_pval", "sd_log_pvals"])
    l = zip(log_true_pvals, sd_log_pvals)
    w.writerows(l)
    f.close()

    #Plot things....
    pngFile_log_pvals = outputFile + ".rob.log_pval.png"
    pngFile_pval = outputFile + ".rob.pval.png"
    pngFile_sd_log_pval = outputFile + ".rob.sd_log_pval.png"
    pngFile_sd_pval = outputFile + ".rob.sd_pval.png"

    min_val = min(true_pvals)
    max_val = max(true_pvals)
    val_range = max_val - min_val

    min_log_val = min(log_true_pvals)
    max_log_val = max(log_true_pvals)
    log_val_range = max_val - min_val

    print "Plotting graphs"
    import matplotlib
    matplotlib.use('Agg')
    import matplotlib.pyplot as plt
    plt.figure(figsize=(10, 7))
    max_perm_val = 0
    min_perm_val = 0
    for i in range(0, len(perm_pvalues_list)):
        delta_log_pvals = delta_log_pvals_list[i]
        plt.plot(log_true_pvals, delta_log_pvals, "b.")
        max_perm_val = max(max_perm_val, max(delta_log_pvals))
        min_perm_val = min(min_perm_val, min(delta_log_pvals))
    perm_val_range = max_perm_val - min_perm_val
    v = [
        min_log_val - 0.02 * log_val_range, max_log_val + 0.02 * log_val_range,
        min_perm_val - 0.02 * perm_val_range,
        max_perm_val + 0.02 * perm_val_range
    ]
    plt.axis(v)
    plt.savefig(pngFile_log_pvals, format="png")

    plt.figure(figsize=(10, 7))
    max_perm_val = 0
    min_perm_val = 0
    for i in range(0, len(perm_pvalues_list)):
        delta_pvals = delta_pvals_list[i]
        plt.plot(true_pvals, delta_pvals, "b.")
        max_perm_val = max(max_perm_val, max(delta_pvals))
        min_perm_val = min(min_perm_val, min(delta_pvals))
    perm_val_range = max_perm_val - min_perm_val
    plt.axis([
        min_val - 0.02 * val_range, max_val + 0.02 * val_range,
        min_perm_val - 0.02 * perm_val_range,
        max_perm_val + 0.02 * perm_val_range
    ])
    plt.savefig(pngFile_pval, format="png")

    plt.figure(figsize=(10, 7))
    max_sd_log_pval = max(sd_log_pvals)
    min_sd_log_pval = min(sd_log_pvals)
    sd_val_range = max_sd_log_pval - min_sd_log_pval
    plt.plot(log_true_pvals, sd_log_pvals, "b.")
    plt.axis([
        min_log_val - 0.02 * log_val_range, max_log_val + 0.02 * log_val_range,
        min_sd_log_pval - 0.02 * sd_val_range,
        max_sd_log_pval + 0.02 * sd_val_range
    ])
    plt.savefig(pngFile_sd_log_pval, format="png")

    plt.figure(figsize=(10, 7))
    max_sd_pval = max(sd_pvals)
    min_sd_pval = min(sd_pvals)
    sd_val_range = max_sd_pval - min_sd_pval
    plt.plot(true_pvals, sd_pvals, "b.")
    plt.axis([
        min_val - 0.02 * val_range, max_val + 0.02 * val_range,
        min_sd_pval - 0.02 * sd_val_range, max_sd_pval + 0.02 * sd_val_range
    ])
    plt.savefig(pngFile_sd_pval, format="png")
    print "Done testing robustness"
Beispiel #2
0
def _robustness_test_(all_snps,phenVals,outputFile,filter=0.1):
	import analyzePhenotype
	import analyzeHaplotype
	import random
	
	new_all_snps = []
	for snp in all_snps:
		if snp.count(0)>1 and snp.count(1)>1:
			new_all_snps.append(snp)
	print "Filtered",len(all_snps)-len(new_all_snps)," with minor allele count <2."
	all_snps = new_all_snps
	

	def getLeaveOneOutK(K,i):
		l = range(0,len(K))
		l.pop(i)
		new_k = numpy.core.take(K,l,0)
		new_k = numpy.core.take(new_k,l,1)
		return new_k
	
	print "Calculating kinship"
	t1 = time.time()
	K = calcKinship(all_snps)
	t2 = time.time()
	print "Took",t2-t1,"seconds."

	
	"""
	Leave one out test..
	"""
	if filter <1.0:
		snps = random.sample(all_snps,int(len(all_snps)*filter))
		print "Number of SNPs:",len(snps)
	else:
		snps = all_snps 
#	K = calcKinship(snps)

	
	print "running EMMA"
	t1 = time.time()
	true_pvals = _runEmma_(snps,phenVals,K)["ps"]
	true_pvals = map(float,true_pvals)
	t2 = time.time()
	print "Took",t2-t1,"seconds."

	
	log_true_pvals = []
	for pval in true_pvals:
		log_true_pvals.append(-math.log(pval,10))
	
	perm_pvalues_list = []
	for i in range(0,len(phenVals)):
		newPhenvals = phenVals[:]
		newPhenvals.pop(i)
				
		newSNPs = []
		for snp in snps:
			newSNP = snp[:]
			newSNP.pop(i)
			newSNPs.append(newSNP)
		
		new_k = getLeaveOneOutK(K,i)
		
		print "running EMMA"
		t1 = time.time()
		pvals = _runEmma_(newSNPs,newPhenvals,new_k)["ps"]
		pvals = map(float,pvals)
		t2 = time.time()
		print "Took",t2-t1,"seconds."

		perm_pvalues_list.append(pvals)
		
		
	
	delta_pvals_list = []
	delta_log_pvals_list = []
	for perm_pvals in perm_pvalues_list:
		log_pvals = []
		delta_pvals = []
		delta_log_pvals = []
		for i in range(0,len(true_pvals)):
			pval = perm_pvals[i]
			true_pval = true_pvals[i]
			delta_pvals.append(true_pval-pval)

			log_true_pval = log_true_pvals[i]
			log_pval = -math.log(pval,10)
			log_pvals.append(log_pval)
			delta_log_pvals.append(log_true_pval-log_pval)
		
		delta_pvals_list.append(delta_pvals)
		delta_log_pvals_list.append(delta_log_pvals)
	
	sd_log_pvals = []
	sd_pvals = []
	t_delta_log_pvals_list = map(list,zip(*delta_log_pvals_list))
	t_delta_pvals_list = map(list,zip(*delta_pvals_list))
	for i in range(0,len(true_pvals)):
		sd_log_pvals.append(util.calcSD(t_delta_log_pvals_list[i]))
		sd_pvals.append(util.calcSD(t_delta_pvals_list[i]))
	
	
	#Write SDs out to file, to be able to replot, or plot together with other methods... etc
	import csv
	sd_log_pval_file = outputFile+".rob.log_pvals_sd"
	f = open(sd_log_pval_file,"w")
	w = csv.writer(f)
	w.writerow(["log_true_pval","sd_log_pvals"])
	l = zip(log_true_pvals,sd_log_pvals)
	w.writerows(l)
	f.close()
	
	
	#Plot things....
	pngFile_log_pvals = outputFile+".rob.log_pval.png"
	pngFile_pval = outputFile+".rob.pval.png"
	pngFile_sd_log_pval = outputFile+".rob.sd_log_pval.png"
	pngFile_sd_pval = outputFile+".rob.sd_pval.png"


	min_val = min(true_pvals)
	max_val = max(true_pvals)
	val_range = max_val-min_val
	
	min_log_val = min(log_true_pvals)
	max_log_val = max(log_true_pvals)
	log_val_range = max_val-min_val


	print "Plotting graphs"
	import matplotlib
	matplotlib.use('Agg')
	import matplotlib.pyplot as plt
	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_log_pvals = delta_log_pvals_list[i]
		plt.plot(log_true_pvals,delta_log_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_log_pvals))
		min_perm_val = min(min_perm_val,min(delta_log_pvals))
	perm_val_range = max_perm_val - min_perm_val
	v = [min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range]
	plt.axis(v)
	plt.savefig(pngFile_log_pvals, format = "png")

	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_pvals = delta_pvals_list[i]
		plt.plot(true_pvals,delta_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_pvals))
		min_perm_val = min(min_perm_val,min(delta_pvals))
	perm_val_range = max_perm_val - min_perm_val
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range])
	plt.savefig(pngFile_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_log_pval = max(sd_log_pvals)
	min_sd_log_pval = min(sd_log_pvals)
	sd_val_range = max_sd_log_pval-min_sd_log_pval
	plt.plot(log_true_pvals,sd_log_pvals,"b.")
	plt.axis([min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_sd_log_pval-0.02*sd_val_range, max_sd_log_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_log_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_pval = max(sd_pvals)
	min_sd_pval = min(sd_pvals)
	sd_val_range = max_sd_pval-min_sd_pval
	plt.plot(true_pvals,sd_pvals,"b.")
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_sd_pval-0.02*sd_val_range, max_sd_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_pval, format = "png")
	print "Done testing robustness"
Beispiel #3
0
def _robustness_test_(all_snps,phenVals,outputFile,filter=0.1,test_type = "KW",):
	"""
	Leave one out test..
	"""
	
	new_all_snps = []
	for snp in all_snps:
		if snp.count(0)>1 and snp.count(1)>1:
			new_all_snps.append(snp)
	print "Filtered",len(all_snps)-len(new_all_snps)," with minor allele count <2."
	all_snps = new_all_snps

	if filter <1.0:
		snps = random.sample(all_snps,int(len(all_snps)*filter))
		print "Number of SNPs:",len(snps)
	else:
		snps = all_snps 

	if test_type=="KW":
		print "running KW"
		t1 = time.time()
		true_pvals = util.kruskal_wallis(snps,phenVals)["ps"]
		t2 = time.time()
		print "Took",t2-t1,"seconds."
	elif test_type=="Fisher":
		print "running Fisher's exact test"
		t1 = time.time()
		true_pvals = run_fet(snps,phenVals)
		t2 = time.time()
		print "Took",t2-t1,"seconds."
	
	log_true_pvals = []
	for pval in true_pvals:
		log_true_pvals.append(-math.log(pval,10))
	
	perm_pvalues_list = []
	for i in range(0,len(phenVals)):
		newPhenvals = phenVals[:]
		newPhenvals.pop(i)
				
		newSNPs = []
		for snp in snps:
			newSNP = snp[:]
			newSNP.pop(i)
			newSNPs.append(newSNP)
		
		print i
		if test_type=="KW":
			print "running KW"
			t1 = time.time()
			pvals = util.kruskal_wallis(newSNPs,newPhenvals)["ps"]
			t2 = time.time()
			print "Took",t2-t1,"seconds."
		elif test_type=="Fisher":
			print "running Fisher's exact test"
			t1 = time.time()
			pvals = run_fet(newSNPs,newPhenvals)
			t2 = time.time()
			print "Took",t2-t1,"seconds."
		perm_pvalues_list.append(pvals)
		
		
	
	delta_pvals_list = []
	delta_log_pvals_list = []
	for perm_pvals in perm_pvalues_list:
		log_pvals = []
		delta_pvals = []
		delta_log_pvals = []
		for i in range(0,len(true_pvals)):
			pval = perm_pvals[i]
			true_pval = true_pvals[i]
			delta_pvals.append(true_pval-pval)

			log_true_pval = log_true_pvals[i]
			if pval > 0.0:
				log_pval = -math.log(pval,10)
			else:
				print "Damn those random 0 prob. events: event #", i
				log_pval = -math.log(true_pval,10)
				
			log_pvals.append(log_pval)
			delta_log_pvals.append(log_true_pval-log_pval)
		
		delta_pvals_list.append(delta_pvals)
		delta_log_pvals_list.append(delta_log_pvals)
	
	sd_log_pvals = []
	sd_pvals = []
	t_delta_log_pvals_list = map(list,zip(*delta_log_pvals_list))
	t_delta_pvals_list = map(list,zip(*delta_pvals_list))
	for i in range(0,len(true_pvals)):
		sd_log_pvals.append(util.calcSD(t_delta_log_pvals_list[i]))
		sd_pvals.append(util.calcSD(t_delta_pvals_list[i]))
	
	
	#Write SDs out to file, to be able to replot, or plot together with other methods... etc
	import csv
	sd_log_pval_file = outputFile+".rob.log_pvals_sd"
	f = open(sd_log_pval_file,"w")
	w = csv.writer(f)
	w.writerow(["log_true_pval","sd_log_pvals"])
	l = zip(log_true_pvals,sd_log_pvals)
	w.writerows(l)
	f.close()

	
	#Plot things....
	pngFile_log_pvals = outputFile+".rob.log_pval.png"
	pngFile_pval = outputFile+".rob.pval.png"
	pngFile_sd_log_pval = outputFile+".rob.sd_log_pval.png"
	pngFile_sd_pval = outputFile+".rob.sd_pval.png"


	min_val = min(true_pvals)
	max_val = max(true_pvals)
	val_range = max_val-min_val

	min_log_val = min(log_true_pvals)
	max_log_val = max(log_true_pvals)
	log_val_range = max_val-min_val


	import matplotlib
	matplotlib.use('Agg')
	import matplotlib.pyplot as plt
	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_log_pvals = delta_log_pvals_list[i]
		plt.plot(log_true_pvals,delta_log_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_log_pvals))
		min_perm_val = min(min_perm_val,min(delta_log_pvals))
	perm_val_range = max_perm_val - min_perm_val
	plt.axis([min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range])
	plt.savefig(pngFile_log_pvals, format = "png")

	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_pvals = delta_pvals_list[i]
		plt.plot(true_pvals,delta_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_pvals))
		min_perm_val = min(min_perm_val,min(delta_pvals))
	perm_val_range = max_perm_val - min_perm_val
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range])
	plt.savefig(pngFile_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_log_pval = max(sd_log_pvals)
	min_sd_log_pval = min(sd_log_pvals)
	sd_val_range = max_sd_log_pval-min_sd_log_pval
	plt.plot(log_true_pvals,sd_log_pvals,"b.")
	plt.axis([min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_sd_log_pval-0.02*sd_val_range, max_sd_log_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_log_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_pval = max(sd_pvals)
	min_sd_pval = min(sd_pvals)
	sd_val_range = max_sd_pval-min_sd_pval
	plt.plot(true_pvals,sd_pvals,"b.")
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_sd_pval-0.02*sd_val_range, max_sd_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_pval, format = "png")
Beispiel #4
0
def _robustness_test_(all_snps,phenVals,outputFile,filter=0.1,test_type = "KW",):
	"""
	Leave one out test..
	"""
	
	new_all_snps = []
	for snp in all_snps:
		if snp.count(0)>1 and snp.count(1)>1:
			new_all_snps.append(snp)
	print "Filtered",len(all_snps)-len(new_all_snps)," with minor allele count <2."
	all_snps = new_all_snps

	if filter <1.0:
		snps = random.sample(all_snps,int(len(all_snps)*filter))
		print "Number of SNPs:",len(snps)
	else:
		snps = all_snps 

	if test_type=="KW":
		print "running KW"
		t1 = time.time()
		true_pvals = util.kruskal_wallis(snps,phenVals)["ps"]
		t2 = time.time()
		print "Took",t2-t1,"seconds."
	elif test_type=="Fisher":
		print "running Fisher's exact test"
		t1 = time.time()
		true_pvals = run_fet(snps,phenVals)
		t2 = time.time()
		print "Took",t2-t1,"seconds."
	
	log_true_pvals = []
	for pval in true_pvals:
		log_true_pvals.append(-math.log(pval,10))
	
	perm_pvalues_list = []
	for i in range(0,len(phenVals)):
		newPhenvals = phenVals[:]
		newPhenvals.pop(i)
				
		newSNPs = []
		for snp in snps:
			newSNP = snp[:]
			newSNP.pop(i)
			newSNPs.append(newSNP)
		
		print i
		if test_type=="KW":
			print "running KW"
			t1 = time.time()
			pvals = util.kruskal_wallis(newSNPs,newPhenvals)["ps"]
			t2 = time.time()
			print "Took",t2-t1,"seconds."
		elif test_type=="Fisher":
			print "running Fisher's exact test"
			t1 = time.time()
			pvals = run_fet(newSNPs,newPhenvals)
			t2 = time.time()
			print "Took",t2-t1,"seconds."
		perm_pvalues_list.append(pvals)
		
		
	
	delta_pvals_list = []
	delta_log_pvals_list = []
	for perm_pvals in perm_pvalues_list:
		log_pvals = []
		delta_pvals = []
		delta_log_pvals = []
		for i in range(0,len(true_pvals)):
			pval = perm_pvals[i]
			true_pval = true_pvals[i]
			delta_pvals.append(true_pval-pval)

			log_true_pval = log_true_pvals[i]
			if pval > 0.0:
				log_pval = -math.log(pval,10)
			else:
				print "Damn those random 0 prob. events: event #", i
				log_pval = -math.log(true_pval,10)
				
			log_pvals.append(log_pval)
			delta_log_pvals.append(log_true_pval-log_pval)
		
		delta_pvals_list.append(delta_pvals)
		delta_log_pvals_list.append(delta_log_pvals)
	
	sd_log_pvals = []
	sd_pvals = []
	t_delta_log_pvals_list = map(list,zip(*delta_log_pvals_list))
	t_delta_pvals_list = map(list,zip(*delta_pvals_list))
	for i in range(0,len(true_pvals)):
		sd_log_pvals.append(util.calcSD(t_delta_log_pvals_list[i]))
		sd_pvals.append(util.calcSD(t_delta_pvals_list[i]))
	
	
	#Write SDs out to file, to be able to replot, or plot together with other methods... etc
	import csv
	sd_log_pval_file = outputFile+".rob.log_pvals_sd"
	f = open(sd_log_pval_file,"w")
	w = csv.writer(f)
	w.writerow(["log_true_pval","sd_log_pvals"])
	l = zip(log_true_pvals,sd_log_pvals)
	w.writerows(l)
	f.close()

	
	#Plot things....
	pngFile_log_pvals = outputFile+".rob.log_pval.png"
	pngFile_pval = outputFile+".rob.pval.png"
	pngFile_sd_log_pval = outputFile+".rob.sd_log_pval.png"
	pngFile_sd_pval = outputFile+".rob.sd_pval.png"


	min_val = min(true_pvals)
	max_val = max(true_pvals)
	val_range = max_val-min_val

	min_log_val = min(log_true_pvals)
	max_log_val = max(log_true_pvals)
	log_val_range = max_val-min_val


	import matplotlib
	matplotlib.use('Agg')
	import matplotlib.pyplot as plt
	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_log_pvals = delta_log_pvals_list[i]
		plt.plot(log_true_pvals,delta_log_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_log_pvals))
		min_perm_val = min(min_perm_val,min(delta_log_pvals))
	perm_val_range = max_perm_val - min_perm_val
	plt.axis([min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range])
	plt.savefig(pngFile_log_pvals, format = "png")

	plt.figure(figsize=(10,7))
	max_perm_val = 0
	min_perm_val = 0
	for i in range(0,len(perm_pvalues_list)):
		delta_pvals = delta_pvals_list[i]
		plt.plot(true_pvals,delta_pvals,"b.")
		max_perm_val = max(max_perm_val,max(delta_pvals))
		min_perm_val = min(min_perm_val,min(delta_pvals))
	perm_val_range = max_perm_val - min_perm_val
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_perm_val-0.02*perm_val_range, max_perm_val+0.02*perm_val_range])
	plt.savefig(pngFile_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_log_pval = max(sd_log_pvals)
	min_sd_log_pval = min(sd_log_pvals)
	sd_val_range = max_sd_log_pval-min_sd_log_pval
	plt.plot(log_true_pvals,sd_log_pvals,"b.")
	plt.axis([min_log_val-0.02*log_val_range, max_log_val+0.02*log_val_range, min_sd_log_pval-0.02*sd_val_range, max_sd_log_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_log_pval, format = "png")

	plt.figure(figsize=(10,7))
	max_sd_pval = max(sd_pvals)
	min_sd_pval = min(sd_pvals)
	sd_val_range = max_sd_pval-min_sd_pval
	plt.plot(true_pvals,sd_pvals,"b.")
	plt.axis([min_val-0.02*val_range, max_val+0.02*val_range, min_sd_pval-0.02*sd_val_range, max_sd_pval+0.02*sd_val_range])
	plt.savefig(pngFile_sd_pval, format = "png")