mspire 0.2.4 → 0.3.0
Sign up to get free protection for your applications and to get access to all the features.
- data/INSTALL +1 -0
- data/README +25 -0
- data/Rakefile +129 -40
- data/bin/{find_aa_freq.rb → aafreqs.rb} +2 -2
- data/bin/bioworks_to_pepxml.rb +1 -0
- data/bin/fasta_shaker.rb +1 -96
- data/bin/filter_and_validate.rb +5 -0
- data/bin/{mzxml_to_lmat.rb → ms_to_lmat.rb} +8 -7
- data/bin/prob_validate.rb +6 -0
- data/bin/raw_to_mzXML.rb +2 -2
- data/bin/srf_group.rb +1 -0
- data/bin/srf_to_sqt.rb +40 -0
- data/changelog.txt +68 -0
- data/lib/align/chams.rb +6 -6
- data/lib/align.rb +4 -3
- data/lib/bsearch.rb +120 -0
- data/lib/fasta.rb +318 -86
- data/lib/group_by.rb +10 -0
- data/lib/index_by.rb +11 -0
- data/lib/merge_deep.rb +21 -0
- data/lib/{spec → ms/converter}/mzxml.rb +77 -109
- data/lib/ms/gradient_program.rb +171 -0
- data/lib/ms/msrun.rb +209 -0
- data/lib/{spec/msrun.rb → ms/msrun_index.rb} +7 -40
- data/lib/ms/parser/mzdata/axml.rb +12 -0
- data/lib/ms/parser/mzdata/dom.rb +160 -0
- data/lib/ms/parser/mzdata/libxml.rb +7 -0
- data/lib/ms/parser/mzdata.rb +25 -0
- data/lib/ms/parser/mzxml/axml.rb +11 -0
- data/lib/ms/parser/mzxml/dom.rb +159 -0
- data/lib/ms/parser/mzxml/hpricot.rb +253 -0
- data/lib/ms/parser/mzxml/libxml.rb +15 -0
- data/lib/ms/parser/mzxml/regexp.rb +122 -0
- data/lib/ms/parser/mzxml/rexml.rb +72 -0
- data/lib/ms/parser/mzxml/xmlparser.rb +248 -0
- data/lib/ms/parser/mzxml.rb +175 -0
- data/lib/ms/parser.rb +108 -0
- data/lib/ms/precursor.rb +10 -0
- data/lib/ms/scan.rb +81 -0
- data/lib/ms/spectrum.rb +193 -0
- data/lib/ms.rb +10 -0
- data/lib/mspire.rb +4 -0
- data/lib/roc.rb +61 -1
- data/lib/sample_enzyme.rb +31 -8
- data/lib/scan_i.rb +21 -0
- data/lib/spec_id/aa_freqs.rb +7 -3
- data/lib/spec_id/bioworks.rb +20 -14
- data/lib/spec_id/digestor.rb +139 -0
- data/lib/spec_id/mass.rb +116 -0
- data/lib/spec_id/parser/proph.rb +236 -0
- data/lib/spec_id/precision/filter/cmdline.rb +209 -0
- data/lib/spec_id/precision/filter/interactive.rb +134 -0
- data/lib/spec_id/precision/filter/output.rb +147 -0
- data/lib/spec_id/precision/filter.rb +623 -0
- data/lib/spec_id/precision/output.rb +60 -0
- data/lib/spec_id/precision/prob/cmdline.rb +139 -0
- data/lib/spec_id/precision/prob/output.rb +88 -0
- data/lib/spec_id/precision/prob.rb +171 -0
- data/lib/spec_id/proph/pep_summary.rb +92 -0
- data/lib/spec_id/proph/prot_summary.rb +484 -0
- data/lib/spec_id/proph.rb +2 -466
- data/lib/spec_id/protein_summary.rb +2 -2
- data/lib/spec_id/sequest/params.rb +316 -0
- data/lib/spec_id/sequest/pepxml.rb +1513 -0
- data/lib/spec_id/sequest.rb +2 -1672
- data/lib/spec_id/srf.rb +445 -177
- data/lib/spec_id.rb +183 -95
- data/lib/spec_id_xml.rb +8 -10
- data/lib/transmem/phobius.rb +147 -0
- data/lib/transmem/toppred.rb +368 -0
- data/lib/transmem.rb +157 -0
- data/lib/validator/aa.rb +135 -0
- data/lib/validator/background.rb +73 -0
- data/lib/validator/bias.rb +95 -0
- data/lib/validator/cmdline.rb +260 -0
- data/lib/validator/decoy.rb +94 -0
- data/lib/validator/digestion_based.rb +69 -0
- data/lib/validator/probability.rb +48 -0
- data/lib/validator/prot_from_pep.rb +234 -0
- data/lib/validator/transmem.rb +272 -0
- data/lib/validator/true_pos.rb +46 -0
- data/lib/validator.rb +214 -0
- data/lib/xml.rb +38 -0
- data/lib/xml_style_parser.rb +105 -0
- data/lib/xmlparser_wrapper.rb +19 -0
- data/script/compile_and_plot_smriti_final.rb +97 -0
- data/script/extract_gradient_programs.rb +56 -0
- data/script/get_apex_values_rexml.rb +44 -0
- data/script/mzXML2timeIndex.rb +1 -1
- data/script/smriti_final_analysis.rb +103 -0
- data/script/toppred_to_yaml.rb +47 -0
- data/script/tpp_installer.rb +1 -1
- data/{test/tc_align.rb → specs/align_spec.rb} +21 -27
- data/{test/tc_bioworks_to_pepxml.rb → specs/bin/bioworks_to_pepxml_spec.rb} +25 -41
- data/specs/bin/fasta_shaker_spec.rb +259 -0
- data/specs/bin/filter_and_validate__multiple_vals_helper.yaml +202 -0
- data/specs/bin/filter_and_validate_spec.rb +124 -0
- data/specs/bin/ms_to_lmat_spec.rb +34 -0
- data/specs/bin/prob_validate_spec.rb +62 -0
- data/specs/bin/protein_summary_spec.rb +10 -0
- data/{test/tc_fasta.rb → specs/fasta_spec.rb} +354 -310
- data/specs/gi_spec.rb +22 -0
- data/specs/load_bin_path.rb +7 -0
- data/specs/merge_deep_spec.rb +13 -0
- data/specs/ms/gradient_program_spec.rb +77 -0
- data/specs/ms/msrun_spec.rb +455 -0
- data/specs/ms/parser_spec.rb +92 -0
- data/specs/ms/spectrum_spec.rb +89 -0
- data/specs/roc_spec.rb +251 -0
- data/specs/rspec_autotest.rb +149 -0
- data/specs/sample_enzyme_spec.rb +41 -0
- data/specs/spec_helper.rb +133 -0
- data/specs/spec_id/aa_freqs_spec.rb +52 -0
- data/{test/tc_bioworks.rb → specs/spec_id/bioworks_spec.rb} +56 -71
- data/specs/spec_id/digestor_spec.rb +75 -0
- data/specs/spec_id/precision/filter/cmdline_spec.rb +20 -0
- data/specs/spec_id/precision/filter/output_spec.rb +31 -0
- data/specs/spec_id/precision/filter_spec.rb +243 -0
- data/specs/spec_id/precision/prob_spec.rb +111 -0
- data/specs/spec_id/precision/prob_spec_helper.rb +0 -0
- data/specs/spec_id/proph/pep_summary_spec.rb +143 -0
- data/{test/tc_proph.rb → specs/spec_id/proph/prot_summary_spec.rb} +52 -32
- data/{test/tc_protein_summary.rb → specs/spec_id/protein_summary_spec.rb} +85 -0
- data/specs/spec_id/sequest/params_spec.rb +68 -0
- data/specs/spec_id/sequest/pepxml_spec.rb +452 -0
- data/specs/spec_id/sqt_spec.rb +138 -0
- data/specs/spec_id/srf_spec.rb +209 -0
- data/specs/spec_id/srf_spec_helper.rb +302 -0
- data/specs/spec_id_helper.rb +33 -0
- data/specs/spec_id_spec.rb +361 -0
- data/specs/spec_id_xml_spec.rb +33 -0
- data/specs/transmem/phobius_spec.rb +423 -0
- data/specs/transmem/toppred_spec.rb +297 -0
- data/specs/transmem_spec.rb +60 -0
- data/specs/transmem_spec_shared.rb +64 -0
- data/specs/validator/aa_spec.rb +107 -0
- data/specs/validator/background_spec.rb +51 -0
- data/specs/validator/bias_spec.rb +146 -0
- data/specs/validator/decoy_spec.rb +51 -0
- data/specs/validator/fasta_helper.rb +26 -0
- data/specs/validator/prot_from_pep_spec.rb +141 -0
- data/specs/validator/transmem_spec.rb +145 -0
- data/specs/validator/true_pos_spec.rb +58 -0
- data/specs/validator_helper.rb +33 -0
- data/specs/xml_spec.rb +12 -0
- data/test_files/000_pepxml18_small.xml +206 -0
- data/test_files/020a.mzXML.timeIndex +4710 -0
- data/test_files/4-03-03_mzXML/000.mzXML.timeIndex +3973 -0
- data/test_files/4-03-03_mzXML/020.mzXML.timeIndex +3872 -0
- data/test_files/4-03-03_small-prot.xml +321 -0
- data/test_files/4-03-03_small.xml +3876 -0
- data/test_files/7MIX_STD_110802_1.sequest_params_fragment.srf +0 -0
- data/test_files/bioworks-3.3_10prots.xml +5999 -0
- data/test_files/bioworks31.params +77 -0
- data/test_files/bioworks32.params +62 -0
- data/test_files/bioworks33.params +63 -0
- data/test_files/bioworks_single_run_small.xml +7237 -0
- data/test_files/bioworks_small.fasta +212 -0
- data/test_files/bioworks_small.params +63 -0
- data/test_files/bioworks_small.phobius +109 -0
- data/test_files/bioworks_small.toppred.out +2847 -0
- data/test_files/bioworks_small.xml +5610 -0
- data/test_files/bioworks_with_INV_small.xml +3753 -0
- data/test_files/bioworks_with_SHUFF_small.xml +2503 -0
- data/test_files/corrupted_900.srf +0 -0
- data/test_files/head_of_7MIX.srf +0 -0
- data/test_files/interact-opd1_mods_small-prot.xml +304 -0
- data/test_files/messups.fasta +297 -0
- data/test_files/opd1/000.my_answer.100lines.xml +101 -0
- data/test_files/opd1/000.tpp_1.2.3.first10.xml +115 -0
- data/test_files/opd1/000.tpp_2.9.2.first10.xml +126 -0
- data/test_files/opd1/000.v2.1.mzXML.timeIndex +3748 -0
- data/test_files/opd1/000_020-prot.png +0 -0
- data/test_files/opd1/000_020_3prots-prot.mod_initprob.xml +62 -0
- data/test_files/opd1/000_020_3prots-prot.xml +62 -0
- data/test_files/opd1/opd1_cat_inv_small-prot.xml +139 -0
- data/test_files/opd1/sequest.3.1.params +77 -0
- data/test_files/opd1/sequest.3.2.params +62 -0
- data/test_files/opd1/twenty_scans.mzXML +418 -0
- data/test_files/opd1/twenty_scans.v2.1.mzXML +382 -0
- data/test_files/opd1/twenty_scans_answ.lmat +0 -0
- data/test_files/opd1/twenty_scans_answ.lmata +9 -0
- data/test_files/opd1_020_beginning.RAW +0 -0
- data/test_files/opd1_2runs_2mods/interact-opd1_mods__small.xml +753 -0
- data/test_files/orbitrap_mzData/000_cut.xml +1920 -0
- data/test_files/pepproph_small.xml +4691 -0
- data/test_files/phobius.small.noheader.txt +50 -0
- data/test_files/phobius.small.small.txt +53 -0
- data/test_files/s01_anC1_ld020mM.key.txt +25 -0
- data/test_files/s01_anC1_ld020mM.meth +0 -0
- data/test_files/small.fasta +297 -0
- data/test_files/smallraw.RAW +0 -0
- data/test_files/tf_bioworks2excel.bioXML +14340 -0
- data/test_files/tf_bioworks2excel.txt.actual +1035 -0
- data/test_files/toppred.small.out +416 -0
- data/test_files/toppred.xml.out +318 -0
- data/test_files/validator_hits_separate/bias_bioworks_small_HS.fasta +7 -0
- data/test_files/validator_hits_separate/bioworks_small_HS.xml +5651 -0
- data/test_files/yeast_gly_small-prot.xml +265 -0
- data/test_files/yeast_gly_small.1.0_1.0_1.0.parentTimes +6 -0
- data/test_files/yeast_gly_small.xml +3807 -0
- data/test_files/yeast_gly_small2.parentTimes +6 -0
- metadata +273 -57
- data/bin/filter.rb +0 -6
- data/bin/precision.rb +0 -5
- data/lib/spec/mzdata/parser.rb +0 -108
- data/lib/spec/mzdata.rb +0 -48
- data/lib/spec/mzxml/parser.rb +0 -449
- data/lib/spec/scan.rb +0 -55
- data/lib/spec_id/filter.rb +0 -797
- data/lib/spec_id/precision.rb +0 -421
- data/lib/toppred.rb +0 -18
- data/script/filter-peps.rb +0 -164
- data/test/tc_aa_freqs.rb +0 -59
- data/test/tc_fasta_shaker.rb +0 -149
- data/test/tc_filter.rb +0 -203
- data/test/tc_filter_peps.rb +0 -46
- data/test/tc_gi.rb +0 -17
- data/test/tc_id_class_anal.rb +0 -70
- data/test/tc_id_precision.rb +0 -89
- data/test/tc_msrun.rb +0 -88
- data/test/tc_mzxml.rb +0 -88
- data/test/tc_mzxml_to_lmat.rb +0 -36
- data/test/tc_peptide_parent_times.rb +0 -27
- data/test/tc_precision.rb +0 -60
- data/test/tc_roc.rb +0 -166
- data/test/tc_sample_enzyme.rb +0 -32
- data/test/tc_scan.rb +0 -26
- data/test/tc_sequest.rb +0 -336
- data/test/tc_spec.rb +0 -78
- data/test/tc_spec_id.rb +0 -201
- data/test/tc_spec_id_xml.rb +0 -36
- data/test/tc_srf.rb +0 -262
@@ -0,0 +1,234 @@
|
|
1
|
+
require 'validator'
|
2
|
+
|
3
|
+
require 'set'
|
4
|
+
require 'group_by'
|
5
|
+
require 'shuffle'
|
6
|
+
|
7
|
+
# calculates protein hit precision based on peptide precision
|
8
|
+
class Validator::ProtFromPep < Validator
|
9
|
+
|
10
|
+
# calculate protein precision based on the number of false peptides
|
11
|
+
# returns the precision based on the number of proteins *completely false*
|
12
|
+
# calculates the worst precision by assuming that proteins with the fewest
|
13
|
+
# peptides are all false (before prots with more pephits)
|
14
|
+
# note that this approaches the worst, but is not guaranteed to be worst
|
15
|
+
# unless each pephit maps to a single protein hit.
|
16
|
+
# [worst, normal_mean, normal_stddev]
|
17
|
+
# options
|
18
|
+
# :num_its_normal => Integer, # num iterations for normal (d: 10)
|
19
|
+
# :num_its_worstcase => Integer, # num iterations for worstcase (d: 10)
|
20
|
+
#
|
21
|
+
def prothit_precision(peps, num_false_pephits, opts={})
|
22
|
+
opts[:num_its_normal] ||= 10
|
23
|
+
opts[:num_its_worstcase] ||= 10
|
24
|
+
# get the num_peps_per_protein array
|
25
|
+
worst = worstcase_prothit_precision(peps, num_false_pephits, :num_its => opts[:num_its_worstcase])
|
26
|
+
(normal_mean, normal_stdev) = normal_prothit_precision( peps, num_false_pephits, :num_its => opts[:num_its_normal])
|
27
|
+
[worst, normal_mean, normal_stdev]
|
28
|
+
end
|
29
|
+
|
30
|
+
# returns an array of the number of peptide hits in each protein
|
31
|
+
def num_peps_per_protein(peps)
|
32
|
+
num_pephits_by_prot = Hash.new { 0 }
|
33
|
+
peps.each do |pep|
|
34
|
+
pep.prots.each do |prot|
|
35
|
+
num_pephits_by_prot[prot.reference] += 1
|
36
|
+
end
|
37
|
+
end
|
38
|
+
num_pephits_by_prot.values
|
39
|
+
end
|
40
|
+
|
41
|
+
# returns the worstcase precision. This assumes that every small protein
|
42
|
+
# with the fewest peptide hits is completely 'filled' with incorrect hits in
|
43
|
+
# preference to any higher hit protein.
|
44
|
+
# Where each peptide hit maps to a single protein, this is guaranteed to be
|
45
|
+
# worst-case. If this doesn't hold, there are some extreme cases where a
|
46
|
+
# poorer precision could be generated, but this is still probably fairly
|
47
|
+
# close. Thus, a slightly different answer may be generated each time.
|
48
|
+
# ...variation is produced by shuffling the order of the proteins from which
|
49
|
+
# peptides are removed within groups of proteins having the same number of
|
50
|
+
# peptides.
|
51
|
+
# This method does NOT require that the prothits be updated to reflect only
|
52
|
+
# those pephits being passed in.
|
53
|
+
#
|
54
|
+
# validator.worstcase_prothit_precision(peps, 14, 1) # => 0.232111
|
55
|
+
#
|
56
|
+
# options:
|
57
|
+
# :num_its => Integer (default: 10) number of times to run (finds minimum)
|
58
|
+
# :one_prot_per_pep => true | *false assumes each peptide maps to a
|
59
|
+
# single protein
|
60
|
+
def worstcase_prothit_precision(peps, num_false_pephits, opts = {})
|
61
|
+
num_its = opts[:num_its] || 10
|
62
|
+
one_prot_per_pep = opts[:one_prot_per_pep] # nil or false still == false
|
63
|
+
one_prot_per_pep = false if one_prot_per_pep == nil
|
64
|
+
|
65
|
+
|
66
|
+
##############################################
|
67
|
+
# The END Cases (can be dealt with quickly)
|
68
|
+
##############################################
|
69
|
+
if num_false_pephits == 0
|
70
|
+
return 1.0
|
71
|
+
elsif num_false_pephits >= peps.size
|
72
|
+
return 0.0
|
73
|
+
end
|
74
|
+
|
75
|
+
if one_prot_per_pep
|
76
|
+
num_peps_per_prot = num_peps_per_protein(peps)
|
77
|
+
return worstcase_prothit_precision_by_numbers(num_peps_per_prot, num_false_pephits)
|
78
|
+
else
|
79
|
+
#####################################
|
80
|
+
# HERE's the basic plan!!
|
81
|
+
#####################################
|
82
|
+
# order the proteins by num peptides
|
83
|
+
# create a set of peptides
|
84
|
+
# delete peptides from the proteins off the set o' peptides (ensuring that
|
85
|
+
# a deleted one cannot be deleted twice)
|
86
|
+
|
87
|
+
#####################################
|
88
|
+
# order the proteins by num peptides
|
89
|
+
# and create a hash that holds the peptides (given here) in those proteins
|
90
|
+
prots_to_peps_here = Hash.new {|h,k| h[k] = [] }
|
91
|
+
prots_to_peps_size = Hash.new { 0 }
|
92
|
+
pep_ids = []
|
93
|
+
pep_ids_to_prot_ids = Hash.new {|h,k| h[k] = [] }
|
94
|
+
peps.each do |pep|
|
95
|
+
#puts pep.prots.size
|
96
|
+
pep.prots.each do |prot|
|
97
|
+
#p prot.reference
|
98
|
+
prots_to_peps_here[prot] << pep
|
99
|
+
prots_to_peps_size[prot] += 1
|
100
|
+
pep_ids << pep
|
101
|
+
pep_ids_to_prot_ids[pep] << prot
|
102
|
+
end
|
103
|
+
end
|
104
|
+
prot_ids_listed_by_peps_size = prots_to_peps_size.keys
|
105
|
+
tot_num_prots = prot_ids_listed_by_peps_size.size
|
106
|
+
|
107
|
+
sample = Array.new(num_its)
|
108
|
+
|
109
|
+
srand( 777 )
|
110
|
+
precision_sample = (0...num_its).to_a.map do
|
111
|
+
num_false_pephits_counter = num_false_pephits
|
112
|
+
# create a set of peptides
|
113
|
+
pep_ids_set = pep_ids.to_set
|
114
|
+
# shuffle the proteins within size groups
|
115
|
+
finished = false
|
116
|
+
prot_ids_listed_by_peps_size.group_by {|prot_id| prots_to_peps_size[prot_id] }.sort.each do |k,group_of_proteins_with_same_pep_size|
|
117
|
+
group_of_proteins_with_same_pep_size.shuffle!
|
118
|
+
group_of_proteins_with_same_pep_size.each do |prot_id|
|
119
|
+
prots_to_peps_here[prot_id].each do |pep_id|
|
120
|
+
if pep_ids_set.include?(pep_id) # if 1
|
121
|
+
# remove a peptide
|
122
|
+
pep_ids_set.delete(pep_id)
|
123
|
+
num_false_pephits_counter -= 1
|
124
|
+
if num_false_pephits_counter == 0 # if 2
|
125
|
+
finished = true
|
126
|
+
end # close if 2
|
127
|
+
end # close if 1
|
128
|
+
break if finished # each pep
|
129
|
+
end
|
130
|
+
break if finished # each prot
|
131
|
+
end
|
132
|
+
break if finished # each group_of_proteins_with_same_pep_size
|
133
|
+
end # each group_of_proteins_with_same_pep_size
|
134
|
+
## Figure out the number of proteins left!
|
135
|
+
proteins_still_around = pep_ids_set.inject(Set.new) {|protset,pep_id| protset.merge( pep_ids_to_prot_ids[pep_id]) }
|
136
|
+
|
137
|
+
proteins_still_around.size.to_f / tot_num_prots
|
138
|
+
end # a sample
|
139
|
+
return precision_sample.min
|
140
|
+
end # FINAL else
|
141
|
+
end
|
142
|
+
|
143
|
+
# returns the precision of the worst possible outcome
|
144
|
+
def worstcase_prothit_precision_by_numbers(num_peps_per_prot, num_false_pephits)
|
145
|
+
completely_false_proteins = 0
|
146
|
+
num_peps_per_prot.sort.each do |num_peps|
|
147
|
+
num_false_pephits -= num_peps
|
148
|
+
if num_false_pephits >= 0
|
149
|
+
completely_false_proteins += 1
|
150
|
+
end
|
151
|
+
if num_false_pephits <= 0
|
152
|
+
break
|
153
|
+
end
|
154
|
+
end
|
155
|
+
num_prots = num_peps_per_prot.size
|
156
|
+
(num_prots - completely_false_proteins).to_f/num_prots
|
157
|
+
end
|
158
|
+
|
159
|
+
# normal as in a standard normal distribution of peptide hits per protein
|
160
|
+
# they are distributed randomly and the precision is assumed to take on a
|
161
|
+
# standard normal distribution.
|
162
|
+
# num_peps_per_protein is an array of the number of peptides per protein hit
|
163
|
+
# (these are the true hits)
|
164
|
+
# assumes that the number follows a gaussian distribution (binomial
|
165
|
+
# distributions tend toward gaussians, I believe, at large N)
|
166
|
+
# returns [mean_precision, stdev_precision]
|
167
|
+
# options:
|
168
|
+
# :num_its => Integer (default: 10)
|
169
|
+
#
|
170
|
+
# if num_iterations is set at 1, then only the precision will be returned
|
171
|
+
# though random, the same seed is always used to start this process, meaning
|
172
|
+
# that the same results will be produced on consecutive attempts.
|
173
|
+
#
|
174
|
+
# validator.normal_prothit_precision(peps, 13, :num_its => 1) # -> 0.95433
|
175
|
+
# validator.normal_prothit_precision(peps, 13, :num_its => 2) # -> [0.92002, 1.2223]
|
176
|
+
def normal_prothit_precision( peps, num_false_pephits, opts={})
|
177
|
+
num_iterations = opts[:num_its] || 10
|
178
|
+
srand( 38272 )
|
179
|
+
|
180
|
+
##############################################
|
181
|
+
# The END Cases (can be dealt with quickly)
|
182
|
+
##############################################
|
183
|
+
if num_false_pephits == 0
|
184
|
+
if num_iterations == 1
|
185
|
+
return 1.0
|
186
|
+
else
|
187
|
+
return [1.0, 0.0]
|
188
|
+
end
|
189
|
+
elsif num_false_pephits >= peps.size
|
190
|
+
if num_iterations == 1
|
191
|
+
return 0.0
|
192
|
+
else
|
193
|
+
return [0.0, 0.0]
|
194
|
+
end
|
195
|
+
end
|
196
|
+
|
197
|
+
##############################################
|
198
|
+
# Everything else:
|
199
|
+
##############################################
|
200
|
+
|
201
|
+
sample = Array.new(num_iterations)
|
202
|
+
base_indices = (0...(peps.size)).to_a
|
203
|
+
### ACUTALLY, I THINK WE WANT TO CREATE AND MERGE!!!!
|
204
|
+
# This would mean that only a single hit would validate the protein
|
205
|
+
# if we are subtracting, then we lose the protein on a single peptide!!!!
|
206
|
+
prot_id_set = peps.inject(Set.new) do |prtset, pep|
|
207
|
+
prtset.merge( pep.prots.map {|prot| prot } )
|
208
|
+
end
|
209
|
+
|
210
|
+
tot_num_prots = prot_id_set.size
|
211
|
+
# could also merge off the good indices
|
212
|
+
# TODO: we should optimize based on how many false pephits given...
|
213
|
+
|
214
|
+
precision_sample = (0...num_iterations).to_a.map do
|
215
|
+
shuffled_indices = base_indices.map
|
216
|
+
shuffled_indices.shuffle!
|
217
|
+
good_indices = shuffled_indices[num_false_pephits..-1]
|
218
|
+
still_remaining = Set.new
|
219
|
+
|
220
|
+
peps.values_at(*good_indices).each do |pep|
|
221
|
+
still_remaining.merge(pep.prots.map {|prot| prot })
|
222
|
+
end
|
223
|
+
still_remaining.size.to_f / tot_num_prots
|
224
|
+
end
|
225
|
+
if num_iterations == 1
|
226
|
+
precision_sample.shift
|
227
|
+
else
|
228
|
+
#puts "PRECISION GROUP: "
|
229
|
+
#p precision_sample
|
230
|
+
sample_stats(precision_sample)
|
231
|
+
end
|
232
|
+
end
|
233
|
+
end
|
234
|
+
|
@@ -0,0 +1,272 @@
|
|
1
|
+
require 'validator'
|
2
|
+
require 'validator/digestion_based'
|
3
|
+
require 'transmem'
|
4
|
+
require 'fasta'
|
5
|
+
require 'spec_id/digestor'
|
6
|
+
require 'spec_id/sequest/params'
|
7
|
+
require 'spec_id/sequest/pepxml'
|
8
|
+
|
9
|
+
|
10
|
+
module Validator::Transmem ; end
|
11
|
+
|
12
|
+
# objects of this class can calculate pephit_precision given an array of
|
13
|
+
# SpecID::Pep objects using the pephit_precision method.
|
14
|
+
class Validator::Transmem::Protein < Validator::DigestionBased
|
15
|
+
include Precision::Calculator
|
16
|
+
|
17
|
+
# a hash keyed by index reference which is true if >= min_num_tms
|
18
|
+
attr_accessor :transmem_by_ti_key
|
19
|
+
attr_accessor :transmem_index
|
20
|
+
|
21
|
+
# min_num_tms: Integer (1...), the min # certain transmembrane segments to
|
22
|
+
# consider the protein a transmembrane protein
|
23
|
+
attr_reader :min_num_tms
|
24
|
+
|
25
|
+
# soluble_fraction: *true/false
|
26
|
+
attr_accessor :soluble_fraction
|
27
|
+
|
28
|
+
# correct_wins: *true/false,
|
29
|
+
# if the peptide is found in some proteins that are transmembrane and some
|
30
|
+
# that are not, then if soluble_fraction==true, this peptide will be
|
31
|
+
# considered non-transmembrane. If soluble_fraction==false, then this
|
32
|
+
# will be considered transmembrane.
|
33
|
+
attr_accessor :correct_wins
|
34
|
+
|
35
|
+
# no_include_tm_peps: false or Float (0.0-1.0), peptides that have a
|
36
|
+
# fraction of amino acids that fall inside transmembrane sequences greater
|
37
|
+
# than or equal to the value of the argument will not be considered in the final
|
38
|
+
# calculation of peptide hit precision. (A transmembrane segment is
|
39
|
+
# likely to have very different properties than the rest of the peptides,
|
40
|
+
# so the assumption of equally flyable peptides is broken unless these are
|
41
|
+
# removed) nil or false will skip this filter. A reasonable value is
|
42
|
+
# probably 0.7.
|
43
|
+
attr_accessor :no_include_tm_peps
|
44
|
+
|
45
|
+
# if nil, then this will be calculated whe pephit_precision is called.
|
46
|
+
attr_accessor :transmem_status_hash
|
47
|
+
|
48
|
+
# the file used (toppred or phobius file)
|
49
|
+
attr_accessor :transmem_file
|
50
|
+
|
51
|
+
DEFAULTS = Validator::DigestionBased::DEFAULTS.merge( { :min_num_tms => 1, :soluble_fraction => true, :correct_wins => true, :no_include_tm_peps => false, :transmem_status_hash => nil} )
|
52
|
+
|
53
|
+
# expects a toppred.out file (see transmem/toppred)
|
54
|
+
# other types of transmembrane predictions)
|
55
|
+
# fasta_obj is a Fasta object.
|
56
|
+
# sequest_params_obj is a Sequest::Params object.
|
57
|
+
# OPTIONS:
|
58
|
+
# (see Validator::Transmem::Protein::DEFAULTS for defaults)
|
59
|
+
#
|
60
|
+
# no_include_tm_peps: *false
|
61
|
+
#
|
62
|
+
# NOTE: if fasta_obj and sequest_params_obj are not passed in then
|
63
|
+
# 'false_to_total_ratio' must be set later.
|
64
|
+
def initialize(a_transmem_file, options={})
|
65
|
+
@transmem_file = a_transmem_file
|
66
|
+
opts = self.class::DEFAULTS.merge(options)
|
67
|
+
|
68
|
+
(@min_num_tms, @soluble_fraction, @correct_wins, @no_include_tm_peps, @background, @transmem_status_hash, @false_to_total_ratio, fasta) = opts.values_at(:min_num_tms, :soluble_fraction, :correct_wins, :no_include_tm_peps, :background, :transmem_status_hash, :false_to_total_ratio, :fasta)
|
69
|
+
|
70
|
+
# fasta object is used to update hte phobius index if given
|
71
|
+
# a hash by reference => true/false (depending on min_num_tms)
|
72
|
+
@transmem_index = TransmemIndex.new(@transmem_file, fasta)
|
73
|
+
@transmem_by_ti_key = create_transmem_by_ti_key_hash(@transmem_index, @min_num_tms)
|
74
|
+
end
|
75
|
+
|
76
|
+
# Designates each protein as transmembrane or not depending on :min_num_tms
|
77
|
+
# The hash is keyed by the TransmemIndex key.
|
78
|
+
def create_transmem_by_ti_key_hash(transmem_index, min_num_tms)
|
79
|
+
_transmem_by_ti_key = {}
|
80
|
+
num_certain_hash = transmem_index.num_certain_index
|
81
|
+
num_certain_hash.each do |id, num_certain|
|
82
|
+
if num_certain >= min_num_tms
|
83
|
+
_transmem_by_ti_key[id] = true
|
84
|
+
else
|
85
|
+
_transmem_by_ti_key[id] = false
|
86
|
+
end
|
87
|
+
end
|
88
|
+
_transmem_by_ti_key
|
89
|
+
end
|
90
|
+
|
91
|
+
# returns a hash where each protein (and peptide if given peps) is indexed
|
92
|
+
# with itself with true/false/nil depending on transmembrane status. If
|
93
|
+
# given peptides, and :no_include_tm_peps is not false, will also set the
|
94
|
+
# attribute for peptides.
|
95
|
+
# the attribute (:no_include_tm_peps)
|
96
|
+
# NOTE: if given a list of peptides, this implementation will not overwrite a
|
97
|
+
# protein if it already has a true/false for transmem. This is so that a
|
98
|
+
# lookup does not have to be performed if the value is already defined as
|
99
|
+
# the assumption is that many peptides will point to the same protein.
|
100
|
+
def create_transmem_status_hash(peps)
|
101
|
+
thash = {}
|
102
|
+
peps.each do |pep|
|
103
|
+
pep.prots.each do |prot|
|
104
|
+
if !thash.key?(prot)
|
105
|
+
#prot.transmem == nil
|
106
|
+
thash[prot] = @transmem_by_ti_key[@transmem_index.reference_to_key(prot.reference)]
|
107
|
+
end
|
108
|
+
end
|
109
|
+
if @no_include_tm_peps
|
110
|
+
thash[pep] = pep_is_transmem?(pep)
|
111
|
+
end
|
112
|
+
end
|
113
|
+
thash
|
114
|
+
end
|
115
|
+
|
116
|
+
# sets the false_to_total_ratio and returns self for chaining.
|
117
|
+
# peps will usually be the peptides created by calling:
|
118
|
+
# peps = Digestor.digest( fasta_obj, sequest_params_obj )
|
119
|
+
def set_false_to_total_ratio(peps)
|
120
|
+
tm_hash = create_transmem_status_hash(peps)
|
121
|
+
(tps, fps) = partition(peps, tm_hash)
|
122
|
+
@false_to_total_ratio = fps.size.to_f / (tps.size + fps.size)
|
123
|
+
self
|
124
|
+
end
|
125
|
+
|
126
|
+
def pephit_precision(peps)
|
127
|
+
if !@transmem_status_hash
|
128
|
+
@transmem_status_hash = create_transmem_status_hash(peps)
|
129
|
+
end
|
130
|
+
super(peps)
|
131
|
+
end
|
132
|
+
|
133
|
+
# regardless of transmembrane status of proteins peptide belongs to, asks
|
134
|
+
# what the avg overlap is with transmembrane sequences.
|
135
|
+
def pep_is_transmem?(pep)
|
136
|
+
prts = pep.prots
|
137
|
+
prts_w_keys = 0
|
138
|
+
sum_of_fractions = 0.0
|
139
|
+
prts.each do |prot|
|
140
|
+
key = @transmem_index.reference_to_key(prot.reference)
|
141
|
+
ans = @transmem_index.avg_overlap(key, pep.aaseq, :fraction)
|
142
|
+
if ans
|
143
|
+
sum_of_fractions += ans
|
144
|
+
prts_w_keys += 1
|
145
|
+
end
|
146
|
+
end
|
147
|
+
if prts_w_keys > 0
|
148
|
+
avg_of_fractions = sum_of_fractions / prts_w_keys
|
149
|
+
avg_of_fractions >= @no_include_tm_peps
|
150
|
+
else
|
151
|
+
nil
|
152
|
+
end
|
153
|
+
end
|
154
|
+
|
155
|
+
# each peptide must have prots and the prots must respond true/false to
|
156
|
+
# the 'transmem' method
|
157
|
+
# if given a hash, it will override the @transmem_status_hash
|
158
|
+
def partition(peps, transmem_status_hash=nil)
|
159
|
+
# The fast way to do this is to play with the logic
|
160
|
+
# For the insoluble fraction we calculate as if incorrect wins
|
161
|
+
# and swap the tp's and fp's (I've verified that this is correct
|
162
|
+
# empirically)
|
163
|
+
|
164
|
+
# the code could be cleaner here, but efforts to minimize calls in the
|
165
|
+
# inner loops create this structure...
|
166
|
+
tm_hash = transmem_status_hash || @transmem_status_hash
|
167
|
+
|
168
|
+
my_peps =
|
169
|
+
if @no_include_tm_peps
|
170
|
+
# remove all thos peps with fractional overlap >= @no_include
|
171
|
+
# [1,2,3,4].reject {|n| n >= 3} #-> [1, 2]
|
172
|
+
# remove pep.transmem == true and pep.transmem == nil
|
173
|
+
|
174
|
+
if tm_hash
|
175
|
+
peps.reject do |pep|
|
176
|
+
tm_hash[pep] != false
|
177
|
+
end
|
178
|
+
else
|
179
|
+
peps.reject do |pep|
|
180
|
+
pep_is_transmem?(pep) != false
|
181
|
+
end
|
182
|
+
end
|
183
|
+
else
|
184
|
+
peps
|
185
|
+
end
|
186
|
+
cw = @correct_wins
|
187
|
+
sf = @soluble_fraction
|
188
|
+
if !sf
|
189
|
+
cw = !cw
|
190
|
+
end
|
191
|
+
|
192
|
+
tp = []
|
193
|
+
fp = []
|
194
|
+
|
195
|
+
if cw
|
196
|
+
my_peps.each do |pep|
|
197
|
+
one_prot_is_not_transmem = false
|
198
|
+
not_all_nil = false
|
199
|
+
if tm_hash
|
200
|
+
pep.prots.each do |prot|
|
201
|
+
tm_status = tm_hash[prot]
|
202
|
+
if tm_status == false
|
203
|
+
one_prot_is_not_transmem = true
|
204
|
+
break
|
205
|
+
elsif tm_status == true
|
206
|
+
not_all_nil = true
|
207
|
+
end
|
208
|
+
end
|
209
|
+
else
|
210
|
+
pep.prots.each do |prot|
|
211
|
+
tm_status = @transmem_by_ti_key[@transmem_index.reference_to_key(prot.reference)]
|
212
|
+
if tm_status == false
|
213
|
+
one_prot_is_not_transmem = true
|
214
|
+
break
|
215
|
+
elsif tm_status == true
|
216
|
+
not_all_nil = true
|
217
|
+
end
|
218
|
+
end
|
219
|
+
end
|
220
|
+
if one_prot_is_not_transmem
|
221
|
+
tp << pep
|
222
|
+
else
|
223
|
+
if not_all_nil
|
224
|
+
fp << pep
|
225
|
+
end
|
226
|
+
end
|
227
|
+
end
|
228
|
+
else
|
229
|
+
my_peps.each do |pep|
|
230
|
+
one_prot_is_transmem = false
|
231
|
+
not_all_nil = false
|
232
|
+
if tm_hash
|
233
|
+
pep.prots.each do |prot|
|
234
|
+
tm_status = tm_hash[prot]
|
235
|
+
if tm_status == true
|
236
|
+
one_prot_is_transmem = true
|
237
|
+
break
|
238
|
+
elsif tm_status == false
|
239
|
+
not_all_nil = true
|
240
|
+
end
|
241
|
+
end
|
242
|
+
else
|
243
|
+
pep.prots.each do |prot|
|
244
|
+
tm_status = @transmem_by_ti_key[@transmem_index.reference_to_key(prot.reference)]
|
245
|
+
if tm_status == true
|
246
|
+
one_prot_is_transmem = true
|
247
|
+
break
|
248
|
+
elsif tm_status == false
|
249
|
+
not_all_nil = true
|
250
|
+
end
|
251
|
+
end
|
252
|
+
end
|
253
|
+
if one_prot_is_transmem
|
254
|
+
fp << pep
|
255
|
+
else
|
256
|
+
if not_all_nil
|
257
|
+
tp << pep
|
258
|
+
end
|
259
|
+
end
|
260
|
+
end
|
261
|
+
end
|
262
|
+
if !sf # swap
|
263
|
+
fp,tp = tp,fp
|
264
|
+
cw = !cw
|
265
|
+
end
|
266
|
+
#puts "PARTITION ARRAY"
|
267
|
+
#p [tp, fp].map{|v| v.size}
|
268
|
+
[tp, fp]
|
269
|
+
end
|
270
|
+
|
271
|
+
end
|
272
|
+
|
@@ -0,0 +1,46 @@
|
|
1
|
+
require 'validator'
|
2
|
+
|
3
|
+
class Validator::TruePos < Validator
|
4
|
+
include Precision::Calculator
|
5
|
+
attr_reader :fasta
|
6
|
+
attr_accessor :correct_wins
|
7
|
+
|
8
|
+
# correct_wins means that only a single protein from a pep.aaseq must match
|
9
|
+
# the fasta object for the pep hit to be considered valid. Otherwise, all
|
10
|
+
# must be a match
|
11
|
+
def initialize(fasta_obj, correct_wins = true)
|
12
|
+
@fasta = fasta_obj
|
13
|
+
@fasta_headers = @fasta.prots.map {|prot| prot.header }
|
14
|
+
@correct_wins = correct_wins
|
15
|
+
end
|
16
|
+
|
17
|
+
def partition(peps)
|
18
|
+
if @correct_wins
|
19
|
+
peps.partition do |pep|
|
20
|
+
@fasta_headers.any? do |header|
|
21
|
+
pep.prots.any? do |pepprot|
|
22
|
+
header.include? pepprot.reference
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
else
|
27
|
+
peps.partition do |pep|
|
28
|
+
pep.prots.all? do |pepprot|
|
29
|
+
@fasta_headers.any? do |header|
|
30
|
+
header.include? pepprot.reference
|
31
|
+
end
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
35
|
+
end
|
36
|
+
|
37
|
+
def pephit_precision(peps)
|
38
|
+
(tp, fp) = partition(peps)
|
39
|
+
calc_precision(tp.size, fp.size)
|
40
|
+
end
|
41
|
+
|
42
|
+
def to_param_string
|
43
|
+
"true_positives(tps)=" + ["{fasta=#{@fasta.filename}", "correct_wins=#{@correct_wins}}"].join(", ")
|
44
|
+
end
|
45
|
+
|
46
|
+
end
|