-
Notifications
You must be signed in to change notification settings - Fork 0
/
sad.py
executable file
·239 lines (197 loc) · 9.78 KB
/
sad.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)
import pandas as pd
import numpy as np
import sys
import time, sys, traceback, argparse
from pysnptools.snpreader import Bed
from pysnptools.standardizer import Unit
from estimation import *
################################################################################################
#This script is used to perform SAD variance decomposition in polygenic risk scores.
#The following inputs are necessary in order to perform SAD analysis:
# - A genetic dataset for the target cohort (the prediction cohort where PCA is performed)
# - A file containing summary statistics from a population level GWAS
# - A file containing summary statistics form a sibling level GWAS
# - A p-value threshold to determine which variants are used in the SAD decomposition
# - The set of effect sizes in which to perform the ascertainment ('gwas' and 'sibs')
# - A path to the directory where out files should be written
#
#IMPORTANT: the datasets with effect sizes (gwas and sib) need to be pre-processed to meet a few criteria.
# In particular, a column called "POS" that identifies loci by chromosome and position in GRCh37 coordinates.
# All effect sizes need to be in the same units (e.g. all in cm for a length measurement, or all in SD units)
# The same data on the same, in the same order, for a set of SNPs
# Standard error of effect-size estimate named 'SE'
# P-value for marker named 'PVAL'
# Effect sizes polarized consistently across datasets (i.e. same alternate allele)
# Effect-size estimate named 'BETA'
#
#This script then reads in the datasets, lines up the markers, and estimates
# variance components along each principal component. Also returns sampling variation
# via bootstrapping over Pickrell's ~1700 blocks and significance testing via #sign-flipping in each block.
################################################################################################
__version__ = '1.0.0'
MASTHEAD = "*********************************************************************\n"
MASTHEAD += "* SAD variance decomposition for polygenic scores\n"
MASTHEAD += "* Version {V}\n".format(V=__version__)
MASTHEAD += "* (C) 2024 Samuel Pattillo Smith, Doc Edge, and Arbel Harpak\n"
MASTHEAD += "* University of Texas and University of Southern California\n"
MASTHEAD += "* GNU General Public License v3\n"
MASTHEAD += "*********************************************************************\n"
def sec_to_str(t):
'''Convert seconds to days:hours:minutes:seconds'''
[d, h, m, s, n] = reduce(lambda ll, b : divmod(ll[0], b) + ll[1:], [(t, 1), 60, 60, 24])
f = ''
if d > 0:
f += '{D}d:'.format(D=d)
if h > 0:
f += '{H}h:'.format(H=h)
if m > 0:
f += '{M}m:'.format(M=m)
f += '{S}s'.format(S=s)
return f
class Logger(object):
def __init__(self, fh):
self.log_fh = open(fh, 'w')
def log(self, msg):
self.log_fh.write(msg)
#read in necessary arguments and filepaths, save as appropriate variables
parser = argparse.ArgumentParser()
parser.add_argument("--genetic-file", type=str, default = './', dest = 'genetic_file')
parser.add_argument("--pop-gwas-file", type=str, default = './', dest = 'popgwas')
parser.add_argument("--sib-gwas-file", type=str, default = './', dest = 'sibgwas')
parser.add_argument("--pvalue", type=float, default = 1, dest = 'pval')
parser.add_argument("--ascertainment-set", type=str, default = 'gwas', dest = 'ascertainment')
parser.add_argument("--chrom", type=str, default = 'CHR', dest = 'CHR')
parser.add_argument("--pos", type=str, default = 'POS', dest = 'POS')
parser.add_argument("--chrom-pos", type=str, default = 'chrom.pos', dest = 'chr_pos')
parser.add_argument("--pop-effect", type=str, default = 'BETA', dest = 'POPBETA')
parser.add_argument("--pop-se", type=str, default = 'se', dest = 'POPSE')
parser.add_argument("--sib-effect", type=str, default = 'BETA', dest = 'SIBBETA')
parser.add_argument("--sib-se", type=str, default = 'se', dest = 'SIBSE')
parser.add_argument("--pval-col", type=str, default = 'P', dest = 'P')
parser.add_argument("--pop-pval-col", type=str, default = 'P', dest = 'POP_P')
parser.add_argument('--nboots', type=int, dest = 'nboots', default = 100)
parser.add_argument("--eigenvals", type=str, default = None, dest = 'eigenvalues')
parser.add_argument("--eigenvecs", type=str, default = None, dest = 'eigenvecs')
parser.add_argument('--permutation-test', default=False, action=argparse.BooleanOptionalAction, dest = 'block_perm')
parser.add_argument('--perm-pcs', type=int, dest = 'pcs_to_test', default = 100)
parser.add_argument('--nperm', type = int, dest = 'nperm', default = 1000)
parser.add_argument("--outfile-label", type=str, default = '', dest = 'outlabel')
parser.add_argument("--out", type=str, default = './', dest = 'outpath')
parser.add_argument('--anc-data',type=str, dest = 'anc_data',default = 'support_files/SNPalleles_1000Genomes_allsites.txt.gz')
args = parser.parse_args()
genetic_file = args.genetic_file
popgwas = args.popgwas
sibgwas = args.sibgwas
pval = args.pval
CHR = args.CHR
POS = args.POS
chr_pos = args.chr_pos
POPBETA = args.POPBETA
POPSE = args.POPSE
SIBBETA = args.SIBBETA
SIBSE = args.SIBSE
POP_P = args.POP_P
P = args.P
anc_data = args.anc_data
ascertainment = args.ascertainment
eigenvalues = args.eigenvalues
eigenvecs = args.eigenvecs
outpath = args.outpath
outlabel = args.outlabel
nboots = args.nboots
block_perm = args.block_perm
pcs_to_test = args.pcs_to_test
nperm = args.nperm
nboots = args.nboots
if __name__ == '__main__':
args = parser.parse_args()
if args.outpath is None:
raise ValueError('--outpath is required.')
log = Logger(args.outpath+'/pval.' + str(pval) + '.log')
defaults = vars(parser.parse_args(''))
opts = vars(args)
non_defaults = [x for x in opts.keys() if opts[x] != defaults[x]]
header = MASTHEAD
header += "Call: \n"
header += './sad.py \\\n'
options = ['--'+x.replace('_','-')+' '+str(opts[x])+' \\' for x in non_defaults]
header += '\n'.join(options).replace('True','').replace('False','')
header = header[0:-1]+'\n'
log.log(header)
log.log('Beginning analysis at {T}\n'.format(T=time.ctime()))
start_time = time.time()
if eigenvalues == None:
log.log('No eigenvalues or eigenvectors were provided. PCA will be performed.\n')
else:
log.log('Eigenvalues and eigenvectors provided. Skipping PCA.\n')
log.log('Reading support files...')
#load in the allele info
anc_data = pd.read_csv(anc_data,sep = '\t', compression = 'gzip')
anc_data['chrom.pos'] = anc_data['SNP'].astype(str)
log.log('Done.\n')
#read in genetic data
log.log('Reading population GWAS statistics...')
#read in the summary statistics for populations and siblings
if '.gz' in popgwas:
popgwas = pd.read_csv(popgwas, sep = '\t', compression = 'gzip')
else:
popgwas = pd.read_csv(popgwas, sep = '\t')
popgwas['SE'] = popgwas[POPSE].astype(float)
log.log('Done.\n')
log.log('Reading sibling GWAS statistics...')
if '.gz' in sibgwas:
sibgwas = pd.read_csv(sibgwas, sep = '\t', compression = 'gzip')
else:
sibgwas = pd.read_csv(sibgwas, sep = '\t')
log.log('Done.\n')
#subset the population gwas results to just the sites that are in the target genotype data
log.log('Reading target sample genotypes...')
genotypes = Bed(genetic_file, count_A1 = True)
loci = ['%g'%(y[0]) + ':' + str(y[2])[:-2] for y in genotypes.pos]
popgwas = popgwas[popgwas[chr_pos].isin(loci)]
sibgwas = sibgwas[sibgwas[chr_pos].isin(loci)].drop_duplicates(keep='first')
overlap_snps = sibgwas.merge(popgwas, on = chr_pos,how = 'inner')[[chr_pos]]
popgwas = popgwas[popgwas[chr_pos].isin(overlap_snps[chr_pos].tolist())].drop_duplicates(subset = chr_pos)
sibgwas = sibgwas[sibgwas[chr_pos].isin(overlap_snps[chr_pos].tolist())].drop_duplicates(subset = chr_pos)
statsnps = popgwas[chr_pos].tolist()
statsnps = genotypes.sid_to_index(statsnps)
pc_genotypes = genotypes.read().val[:,statsnps]
snps_nans = np.unique(np.argwhere(np.isnan(pc_genotypes))[:,1])
pc_genotypes = np.delete(pc_genotypes, snps_nans, 1)
statsnps = np.delete(statsnps,snps_nans,0)
popgwas = popgwas.drop(snps_nans)
popgwas = popgwas.reset_index(drop = True)
sibgwas = sibgwas.drop(snps_nans)
sibgwas = sibgwas.reset_index(drop = True)
block_snps = popgwas[[CHR, POS, POP_P]]
log.log('Done.\n')
if ascertainment == 'sibs':
asc_ps = sibgwas[P]
stat_to_geno_df = sibgwas
elif ascertainment == 'gwas':
asc_ps = popgwas[POP_P]
stat_to_geno_df = popgwas
log.log('Beginning SAD decomposition...')
sad = estimate_components(anc_data, pc_genotypes, popgwas[POPBETA].astype(float), popgwas[POPSE].astype(float), \
sibgwas[SIBBETA].astype(float), sibgwas[SIBSE].astype(float), block_snps, asc_ps.astype(float), \
pval, outpath, outlabel, POS, pc_lower_bound=100, eigenvecs= eigenvecs, eigenvalues = eigenvalues, \
boot_se = nboots, block_perm = block_perm, pcs_to_test = pcs_to_test, nperm = nperm)
log.log('Done.\n')
final = sad.outputs()
log.log('Alpha estimate: ' + str(final['alpha']) + ' (' + str(final['alpha_se']) + ')\n')
log.log('nSNPs passing ascertainment filter: ' + str(final['nsnp'])+'\n')
if outlabel != '':
newfile = open(outpath + '/' + outlabel + '.pval.' + str(pval) + '.alpha.txt','w')
alphasefile = open(outpath + '/' + outlabel + '.pval.' + str(pval) + '.alpha.se.txt','w')
nsnpfile = open(outpath + '/' + outlabel + '.pval.' + str(pval) + '.nsnp.txt','w')
final['var_totals'].to_csv(outpath + '/' + outlabel + '.pval.' + str(pval) + '.variance.totals.txt', sep = '\t')
else:
newfile = open(outpath + '/pval.' + str(pval) + '.alpha.txt','w')
nsnpfile = open(outpath + '/pval.' + str(pval) + '.nsnp.txt','w')
final['var_totals'].to_csv(outpath + '/pval.' + str(pval) + '.variance.totals.txt', sep = '\t')
newfile.write(str(final['alpha']))
alphasefile.write(str(final['alpha_se']))
nsnpfile.write(str(final['nsnp']))
log.log('SAD decomposition complete.\n')