def estimate_Jarne2006(self, data_matrix):
"""
2007-08-14
method based on Jarne2006
"""
sys.stderr.write("Jarne2006 method ...\n")
s_estimate_result_instance = s_estimate_result()
from EstimateSelfingGeneration import EstimateSelfingGeneration
EstimateSelfingGeneration_instance = EstimateSelfingGeneration()
locus_allele_prob_vector = EstimateSelfingGeneration_instance.cal_locus_allele_prob_vector(data_matrix)
observed_heterozygosity_vector = self.cal_observed_heterozygosity_vector(data_matrix)
FIS_vector = self.cal_FIS_vector(locus_allele_prob_vector, observed_heterozygosity_vector)
selfing_rate_vector = self.cal_selfing_rate_vector(FIS_vector)
s_estimate_result_instance.FIS_vector = FIS_vector
s_estimate_result_instance.selfing_rate_vector = selfing_rate_vector
return s_estimate_result_instance
def estimate_Jarne2006(self, data_matrix):
"""
2007-08-14
method based on Jarne2006
"""
sys.stderr.write("Jarne2006 method ...\n")
s_estimate_result_instance = s_estimate_result()
from EstimateSelfingGeneration import EstimateSelfingGeneration
EstimateSelfingGeneration_instance = EstimateSelfingGeneration()
locus_allele_prob_vector = EstimateSelfingGeneration_instance.cal_locus_allele_prob_vector(
data_matrix)
observed_heterozygosity_vector = self.cal_observed_heterozygosity_vector(
data_matrix)
FIS_vector = self.cal_FIS_vector(locus_allele_prob_vector,
observed_heterozygosity_vector)
selfing_rate_vector = self.cal_selfing_rate_vector(FIS_vector)
s_estimate_result_instance.FIS_vector = FIS_vector
s_estimate_result_instance.selfing_rate_vector = selfing_rate_vector
return s_estimate_result_instance
def estimate_David2007_g2(self, data_matrix):
"""
2007-08-15
David2007
"""
sys.stderr.write("David2007 g2 method ...\n")
no_of_strains, no_of_snps = data_matrix.shape
s_estimate_result_instance = s_estimate_result()
Hijk = Numeric.zeros([no_of_snps, no_of_snps],
Numeric.Float) #each entry is \sum_k Hik*Hjk
Hijkl = Numeric.zeros(
[no_of_snps, no_of_snps],
Numeric.Float) #each entry is \sum_k \sum_{l!=k} Hik*Hjl
Mij = Numeric.zeros(
[no_of_snps, no_of_snps],
Numeric.Float) #number of individuals with NA at locus both i an j
Mi = Numeric.zeros(
no_of_snps,
Numeric.Float) #number of individuals with NA at locus i
for i in range(no_of_snps):
for j in range(no_of_snps):
if j == i:
continue
for k in range(no_of_strains):
if data_matrix[k, i] == 0:
Mi[i] += 1 #this one is inflated no_of_snps-1 times due to the inner j loop
if data_matrix[k, j] == 0:
Mij[i, j] += 1
Hik = int(
data_matrix[k, i] > 4) #NA=0 is treated as homozygous
Hjk = int(data_matrix[k, j] > 4)
Hijk[i, j] += (Hik * Hjk)
for l in range(no_of_strains):
if l == k:
continue
Hik = int(data_matrix[k, i] >
4) #NA=0 is treated as homozygous
Hjl = int(data_matrix[l, j] > 4)
Hijkl[i, j] += (Hik * Hjl)
Mi = Mi / (no_of_snps - 1)
numerator = 0.0
denominator = 0.0
for i in range(no_of_snps):
for j in range(no_of_snps):
if j == i:
continue
numerator += 1.0 / (no_of_strains - Mij[i, j]) * Hijk[i, j]
denominator += 1.0 / (no_of_strains * (no_of_strains - 1) -
Mi[i] * Mi[j] + Mij[i, j]) * Hijkl[i, j]
if denominator > 0.0:
g2 = numerator / denominator
s_estimate_result_instance.g2_David2007 = g2
else:
g2 = None
if g2: #g2=0 causes ZeroDivisionError: float division in calculating s
s = (1 + 5 * g2 - math.sqrt(1 + 10 * g2 + 9 * g2 * g2)) / (2 * g2)
FIS = s / (2 - s)
from EstimateSelfingGeneration import EstimateSelfingGeneration
EstimateSelfingGeneration_instance = EstimateSelfingGeneration()
locus_allele_prob_vector = EstimateSelfingGeneration_instance.cal_locus_allele_prob_vector(
data_matrix)
observed_heterozygosity_vector = self.cal_observed_heterozygosity_vector(
data_matrix)
FIS_vector = self.cal_FIS_vector(locus_allele_prob_vector,
observed_heterozygosity_vector)
genotyping_error_func = lambda x: (x - FIS) / (1 - FIS)
genotyping_error_rate_vector = map(genotyping_error_func,
FIS_vector)
s_estimate_result_instance.s_g2_David2007 = s
s_estimate_result_instance.FIS_vector = FIS_vector
s_estimate_result_instance.genotyping_error_rate_vector = genotyping_error_rate_vector
return s_estimate_result_instance
def estimate_David2007_g2(self, data_matrix):
"""
2007-08-15
David2007
"""
sys.stderr.write("David2007 g2 method ...\n")
no_of_strains, no_of_snps = data_matrix.shape
s_estimate_result_instance = s_estimate_result()
Hijk = Numeric.zeros([no_of_snps, no_of_snps], Numeric.Float) #each entry is \sum_k Hik*Hjk
Hijkl = Numeric.zeros([no_of_snps, no_of_snps], Numeric.Float) #each entry is \sum_k \sum_{l!=k} Hik*Hjl
Mij = Numeric.zeros([no_of_snps, no_of_snps], Numeric.Float) #number of individuals with NA at locus both i an j
Mi = Numeric.zeros(no_of_snps, Numeric.Float) #number of individuals with NA at locus i
for i in range(no_of_snps):
for j in range(no_of_snps):
if j==i:
continue
for k in range(no_of_strains):
if data_matrix[k,i]==0:
Mi[i] += 1 #this one is inflated no_of_snps-1 times due to the inner j loop
if data_matrix[k,j]==0:
Mij[i,j] += 1
Hik = int(data_matrix[k,i]>4) #NA=0 is treated as homozygous
Hjk = int(data_matrix[k,j]>4)
Hijk[i,j] += (Hik*Hjk)
for l in range(no_of_strains):
if l==k:
continue
Hik = int(data_matrix[k,i]>4) #NA=0 is treated as homozygous
Hjl = int(data_matrix[l,j]>4)
Hijkl[i,j] += (Hik*Hjl)
Mi = Mi/(no_of_snps-1)
numerator = 0.0
denominator = 0.0
for i in range(no_of_snps):
for j in range(no_of_snps):
if j==i:
continue
numerator += 1.0/(no_of_strains-Mij[i,j])*Hijk[i,j]
denominator += 1.0/(no_of_strains*(no_of_strains-1)-Mi[i]*Mi[j] + Mij[i,j])*Hijkl[i,j]
if denominator>0.0:
g2 = numerator/denominator
s_estimate_result_instance.g2_David2007 = g2
else:
g2=None
if g2: #g2=0 causes ZeroDivisionError: float division in calculating s
s = (1+5*g2-math.sqrt(1+10*g2+9*g2*g2))/(2*g2)
FIS = s/(2-s)
from EstimateSelfingGeneration import EstimateSelfingGeneration
EstimateSelfingGeneration_instance = EstimateSelfingGeneration()
locus_allele_prob_vector = EstimateSelfingGeneration_instance.cal_locus_allele_prob_vector(data_matrix)
observed_heterozygosity_vector = self.cal_observed_heterozygosity_vector(data_matrix)
FIS_vector = self.cal_FIS_vector(locus_allele_prob_vector, observed_heterozygosity_vector)
genotyping_error_func = lambda x: (x-FIS)/(1-FIS)
genotyping_error_rate_vector = map(genotyping_error_func, FIS_vector)
s_estimate_result_instance.s_g2_David2007 = s
s_estimate_result_instance.FIS_vector = FIS_vector
s_estimate_result_instance.genotyping_error_rate_vector = genotyping_error_rate_vector
return s_estimate_result_instance
