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
|
# Try to implement method used in Morishita et al.'s Medaka fish genome
# paper here
from collections import defaultdict, Counter
import os
from math import sqrt
import math
import scipy.stats as s
import array as a
from scipy.optimize import fminbound
from scipy.special import gammaln as gamln
from numpy import log, pi, log10, e, log1p, exp
import numpy as np
from .MultiSiteCommon import MultiSiteCommon
class MedakaLdaEnricher(MultiSiteCommon):
def __init__(self, gbmModel, sequence, rawKinetics, m5Cclassifier):
MultiSiteCommon.__init__(self, gbmModel, sequence, rawKinetics)
models = np.genfromtxt(m5Cclassifier, delimiter=',')
self.fwd_model = models[:, 0]
self.rev_model = models[:, 1]
# write a method to take perSiteResults dictionary in and add a column Ca5C
def useLDAmodel(self, kinetics, pos, model, up, down):
""" Test out LDA model """
print("From use LDA model.\n")
res = np.zeros((up + down + 1, 6))
ind = 0
# range from -down to +up
for offset in range(-down, (up + 1)):
a = pos + offset
std = kinetics[a]["tErr"] * sqrt(kinetics[a]["coverage"])
mErr = 0.01 + 0.03 * \
kinetics[a]["modelPrediction"] + 0.06 * \
kinetics[a]["modelPrediction"] ** 2
den = sqrt(mErr ** 2 + std ** 2)
t0 = (kinetics[a]["tMean"] - kinetics[a]["modelPrediction"]) / den
res[ind, ] = [kinetics[a]["tMean"], kinetics[a]["modelPrediction"],
std, np.exp(t0) - 0.01, kinetics[a]["ipdRatio"], den]
ind += 1
predictors = np.hstack(np.log(res + 0.01).transpose())
tmp = sum(np.multiply(predictors, model[1:])) + model[0]
return tmp
def callLDAstrand(self, kinetics, strand, model, up, down):
print("From callLDAstrand.\n")
tmp = [d for d in kinetics if d["strand"] == strand]
tmp.sort(key=lambda x: x["tpl"])
L = len(tmp)
for pos in range(down, (L - up)):
if (strand == 0 and tmp[pos]["base"] ==
'C' and tmp[pos + 1]["base"] == 'G'):
tmp[pos]["Ca5C"] = self.useLDAmodel(tmp, pos, model, up, down)
if (strand == 1 and tmp[pos]["base"] ==
'C' and tmp[pos - 1]["base"] == 'G'):
tmp[pos -
1]["Ca5C"] = self.useLDAmodel(tmp, pos, model, up, down)
return tmp
def aggregate(self, dataset, group_by_key, sum_value_key):
print("From aggregate.\n")
emp = {}
for item in dataset:
if sum_value_key in item:
if item[group_by_key] in emp:
emp[item[group_by_key]] += item[sum_value_key]
else:
emp[item[group_by_key]] = item[sum_value_key]
# Need to go back over the set again?
for item in dataset:
if sum_value_key in item:
item[sum_value_key] = emp[item[group_by_key]]
return dataset
def callEnricherFunction(self, kinetics, up=10, down=10):
print("From callEnricher function.\n")
fwd = self.callLDAstrand(kinetics, 0, self.fwd_model, up, down)
rev = self.callLDAstrand(kinetics, 1, self.rev_model, up, down)
res = fwd + rev
res.sort(key=lambda x: x["tpl"])
# Would like to (1) find rows where tpl is the same, (2) add the Ca5C columns for these rows,
# (3) store the same result under Ca5C for each row.
# In R, this would be: a <- aggregate(Ca5C ~ tpl, df, sum )
res = self.aggregate(res, 'tpl', 'Ca5C')
return res
|
