def get_data_top_k_independent(input_fn, ks, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
return get_idabs_top_k_independent(input_fn, ks, birth, dobs, expr)
def get_data_top_k_independent(input_fn, ks, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
return get_idabs_top_k_independent(input_fn, ks, birth, dobs, expr)
def snp_sampling(input_fn):
""" Vary the size of subsets of SNPs (10%, 20%, ..., 100%),
and show how identifiability is affected. """
subset_sizes = [10 * (i+1) for i in range(10)]
sub_fn = input_fn[:-4] + '.subset.vcf'
uniqs_dob = []
uniqs = []
dobs = utils.synthetic_dob(2504)
for size in subset_sizes:
print 'Using random', size, '% of SNPs'
utils.make_toy(data_path+input_fn, data_path+sub_fn, size)
uniq_dob, snp_num = snp_info.identifiability(data_path+sub_fn, use_dob=True, dobs=dobs)
uniq, snp_num = snp_info.identifiability(data_path+sub_fn, use_dob=False, dobs=dobs)
uniqs_dob.append(uniq_dob)
uniqs.append(uniq)
print
if config.birth_type == 'date':
dob = 'DOB'
elif config.birth_type == 'year':
dob = 'YOB'
plt.plot(subset_sizes, uniqs, 'b', label='w/o '+dob)
plt.plot(subset_sizes, uniqs_dob, 'r', label='w/ '+dob)
ax = plt.subplot()
d = (subset_sizes[1] - subset_sizes[0]) * 0.1
xlim = [subset_sizes[0] - d, subset_sizes[-1] + d]
ylim = [-0.05, 1.05]
ax.set_xlim(xlim)
ax.set_ylim(ylim)
plt.legend(loc='best', shadow=True)
plt.title('Identifiability using random X % of '+ str(snp_num) +' SNPs')
plt.savefig(plot_path+'snp_sampling_'+input_fn[:-4]+'_'+dob+'.pdf')
plt.close()
def get_data_vary_af_threshold(input_fn, thresholds, birth, expr):
""" Vary allele frequency threshold and plot change in identifiability. """
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
return get_idabs_vary_af_threshold(input_fn, thresholds, birth, expr, dobs)
def get_data_top_k_entropy(input_fn, ks, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
print len(dobs)
return get_idabs_top_k_entropy(input_fn, ks, birth, dobs, expr)
def get_data_vary_rsq(input_fn, rsqs, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
print len(dobs)
return get_idabs_vary_rsq(input_fn, rsqs, birth, dobs, expr)
def get_data_top_k_entropy(input_fn, ks, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
print len(dobs)
return get_idabs_top_k_entropy(input_fn, ks, birth, dobs, expr)
def get_data_vary_af_threshold(input_fn, thresholds, birth, expr):
""" Vary allele frequency threshold and plot change in identifiability. """
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
return get_idabs_vary_af_threshold(input_fn, thresholds, birth, expr, dobs)
def get_data_vary_rsq(input_fn, rsqs, birth, expr):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
print len(dobs)
return get_idabs_vary_rsq(input_fn, rsqs, birth, dobs, expr)
def expr_vary_C(W, C, crange, birth=''):
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
idabs = []
for c in crange:
value, knap_snps = solve_knapsack(W, c)
idab, _ = snp_info.identifiability(input_vcf, True, dobs=dobs, snps=knap_snps)
print 'accuracy at', c, 'th round:', idab
idabs.append(idab)
return idabs
def expr_vary_W(W, C, birth=''):
idabs = []
if birth:
config.birth_type = birth
dobs = utils.synthetic_dob(2504)
else:
dobs = []
for w in range(1, W+1):
value, knap_snps = solve_knapsack(w, C)
idab, _ = snp_info.identifiability(input_vcf, False, dobs=dobs, snps=knap_snps)
print 'accuracy at', w, 'th round:', idab
idabs.append(idab)
return idabs
def make_hash(input_fn, output_fn, use_dob=True, dobs=None):
""" Generate hash code for individuals in input file. """
vcf_reader = vcf.Reader(open(data_path + input_fn, 'rb'))
m = len(vcf_reader.samples)
n = sum(1 for _ in vcf_reader)
if dobs == None:
dobs = utils.synthetic_dob(m)
genotypes = snp_info.genotypes(input_fn, m, n, use_dob, dobs)
hashes = hash_func(genotypes)
with open(data_path + output_fn, 'wb') as outfile:
for hash in hashes:
outfile.write(hash)
outfile.write('\n')
def make_hash(input_fn, output_fn, use_dob=True, dobs=None):
""" Generate hash code for individuals in input file. """
vcf_reader = vcf.Reader(open(data_path+input_fn, 'rb'))
m = len(vcf_reader.samples)
n = sum(1 for _ in vcf_reader)
if dobs == None:
dobs = utils.synthetic_dob(m)
genotypes = snp_info.genotypes(input_fn, m, n, use_dob, dobs)
hashes = hash_func(genotypes)
with open(data_path+output_fn, 'wb') as outfile:
for hash in hashes:
outfile.write(hash)
outfile.write('\n')
def snp_sampling(input_fn):
""" Vary the size of subsets of SNPs (10%, 20%, ..., 100%),
and show how identifiability is affected. """
subset_sizes = [10 * (i + 1) for i in range(10)]
sub_fn = input_fn[:-4] + '.subset.vcf'
uniqs_dob = []
uniqs = []
dobs = utils.synthetic_dob(2504)
for size in subset_sizes:
print 'Using random', size, '% of SNPs'
utils.make_toy(data_path + input_fn, data_path + sub_fn, size)
uniq_dob, snp_num = snp_info.identifiability(data_path + sub_fn,
use_dob=True,
dobs=dobs)
uniq, snp_num = snp_info.identifiability(data_path + sub_fn,
use_dob=False,
dobs=dobs)
uniqs_dob.append(uniq_dob)
uniqs.append(uniq)
print
if config.birth_type == 'date':
dob = 'DOB'
elif config.birth_type == 'year':
dob = 'YOB'
plt.plot(subset_sizes, uniqs, 'b', label='w/o ' + dob)
plt.plot(subset_sizes, uniqs_dob, 'r', label='w/ ' + dob)
ax = plt.subplot()
d = (subset_sizes[1] - subset_sizes[0]) * 0.1
xlim = [subset_sizes[0] - d, subset_sizes[-1] + d]
ylim = [-0.05, 1.05]
ax.set_xlim(xlim)
ax.set_ylim(ylim)
plt.legend(loc='best', shadow=True)
plt.title('Identifiability using random X % of ' + str(snp_num) + ' SNPs')
plt.savefig(plot_path + 'snp_sampling_' + input_fn[:-4] + '_' + dob +
'.pdf')
plt.close()
def test_synthetic_dob(): dobs = utils.synthetic_dob(2504) assert len(dobs) == 2504, "Error in utils.synthetic_dob()"
def test_synthetic_dob():
dobs = utils.synthetic_dob(2504)
assert len(dobs) == 2504, "Error in utils.synthetic_dob()"