+ rearranged code
[qpalma.git] / qpalma / Configuration.py
1 #!/usr/bin/env python
2 # -*- coding: utf-8 -*-
3
4 import numpy.matlib
5 import os.path
6 import cPickle
7
8 ###############################################################################
9 #
10 # Load a random but fixed initial parameter vector this makes debugging easier
11 #
12 ###############################################################################
13
14 fixedParamQ = cPickle.load(open('/fml/ag-raetsch/home/fabio/svn/projects/QPalma/scripts/randInitParam.pickle'))
15
16 ###########################################################
17 #
18 # The parameters for the QPalma algorithm
19 #
20 #
21 C = 1
22
23 min_intron_len = 20
24 max_intron_len = 2000
25
26 min_svm_score = 0.0
27 max_svm_score = 1.0
28
29 numConstraintsPerRound = 100
30
31 ###############################################################################
32 #
33 # CHOOSING THE MODE
34 #
35 # 'normal' means work like Palma 'using_quality_scores' means work like Palma
36 # plus using sequencing quality scores
37 #
38 ###############################################################################
39
40 #mode = 'normal'
41 mode = 'using_quality_scores'
42
43 ###############################################################################
44 #
45 # When using quality scores our scoring function is defined as
46 #
47 # f: S_e x R x S -> R
48 #
49 # where S_e is {A,C,G,T,N} and S = {A,C,G,T,N,-}
50 #
51 # as opposed to a usage without quality scores when we only have
52 #
53 # f: S x S -> R
54 #
55 # The matrix of plifs is defined as follows:
56 #
57 # elem | - a c g t n
58 # -------------------------
59 # idx | 0 1 2 3 4 5
60 #
61 #
62 # dna
63 #
64 # - a c g t n
65 # a
66 # est c
67 # g
68 # t
69 # n
70 #
71 # so the index can be calculated as (estnum-1)*6 + dnanum.
72 #
73 # At ests we do not have gaps with quality scores so we look up the matchmatrix
74 # instead.
75 ###############################################################################
76
77 #numLengthSuppPoints = 30
78 #numDonSuppPoints = 30
79 #numAccSuppPoints = 30
80 #numQualSuppPoints = 16
81
82 numLengthSuppPoints = 20
83 numDonSuppPoints = 20
84 numAccSuppPoints = 20
85 numQualSuppPoints = 8
86
87 min_qual = -1
88 max_qual = 40
89
90 USE_OPT = True
91
92 if mode == 'normal':
93 sizeMatchmatrix = (6,6)
94 estPlifs = 0
95 dnaPlifs = 0
96 numQualPlifs = estPlifs*dnaPlifs
97 elif mode == 'using_quality_scores':
98 sizeMatchmatrix = (6,1)
99 estPlifs = 5
100 dnaPlifs = 6
101 numQualPlifs = estPlifs*dnaPlifs
102 else:
103 assert False, 'Wrong operation mode specified'
104
105 totalQualSuppPoints = numQualPlifs*numQualSuppPoints
106
107 numFeatures = numDonSuppPoints + numAccSuppPoints\
108 + numLengthSuppPoints + sizeMatchmatrix[0]*sizeMatchmatrix[1] + totalQualSuppPoints
109
110
111 ###############################################################################
112 #
113 # GENERAL SETTINGS CONCERNING THE SOLVER
114 #
115 #
116 #
117 ###############################################################################
118
119 iter_steps = 40
120 remove_duplicate_scores = False
121 print_matrix = False
122 anzpath = 2
123
124 if mode == 'normal':
125 fixedParam = fixedParam[:numFeatures]
126 elif mode == 'using_quality_scores':
127 fixedParam = fixedParamQ[:numFeatures]
128 else:
129 assert False, 'Wrong operation mode specified'
130
131 ###############################################################################
132 #
133 # DATA SETTINGS CONCERNING THE SPLITS AND FILE LOCATIONS
134 #
135 #
136 #
137 ###############################################################################
138
139 training_begin = 0
140 training_end = 2000
141
142 prediction_begin = 2000
143 prediction_end = 3500
144
145 joinp = os.path.join
146
147 tmp_dir = '/fml/ag-raetsch/home/fabio/tmp/solexa_tmp'
148 data_path = '/fml/ag-raetsch/share/projects/qpalma/solexa'
149
150 original_path = joinp(data_path,'original_solexa_data')
151 annot_path = joinp(data_path,'annotation_data')
152 remapped_path = joinp(data_path,'remapped_solexa_data')
153
154 dna_flat_fn = '/fml/ag-raetsch/share/projects/genomes/A_thaliana_best/genome/'
155 gff_fn = joinp(annot_path,'TAIR7_GFF3_genes_Chr%s.gff_v1')
156 filtered_fn = joinp(data_path,'allFilteredReads_04_02_2008')
157 remapped_fn = joinp(remapped_path,'map_best_hit.18.unambig')
158
159 dataset_fn = '/fml/ag-raetsch/home/fabio/svn/projects/QPalma/scripts/chr1_dataset.pickle'
160
161 ###############################################################################
162 #
163 # SANITY CHECKS
164 #
165 ###############################################################################
166 assert numQualPlifs >= 0
167 assert numDonSuppPoints > 1
168 assert numAccSuppPoints > 1
169 assert numLengthSuppPoints > 1
170 assert numQualSuppPoints > 1
171
172 assert os.path.exists(dna_flat_fn), 'DNA data does not exist!'
173 assert os.path.exists(filtered_fn), 'EST/Reads data does not exist!'
174 assert os.path.exists(remapped_fn), 'EST/Reads data does not exist!'