big_simon 0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +21 -0
- data/.rspec +3 -0
- data/.travis.yml +7 -0
- data/CODE_OF_CONDUCT.md +74 -0
- data/COPYING +674 -0
- data/Gemfile +6 -0
- data/Gemfile.lock +41 -0
- data/README.md +37 -0
- data/Rakefile +6 -0
- data/big_simon.gemspec +30 -0
- data/bin/console +14 -0
- data/bin/setup +8 -0
- data/lib/big_simon/version.rb +3 -0
- data/lib/big_simon.rb +100 -0
- data/vendor/bin/mac/WIsH +0 -0
- data/vendor/bin/mac/computeMeasure.out +0 -0
- data/vendor/bin/mac/computeMeasure_onlyd2star.out +0 -0
- data/vendor/bin/mac/countKmer.out +0 -0
- data/vendor/bin/mac/vhm.py +381 -0
- metadata +118 -0
data/Gemfile.lock
ADDED
@@ -0,0 +1,41 @@
|
|
1
|
+
PATH
|
2
|
+
remote: .
|
3
|
+
specs:
|
4
|
+
big_simon (0.0.1)
|
5
|
+
rya (~> 0.1.3)
|
6
|
+
|
7
|
+
GEM
|
8
|
+
remote: https://rubygems.org/
|
9
|
+
specs:
|
10
|
+
abort_if (0.2.0)
|
11
|
+
diff-lcs (1.3)
|
12
|
+
rake (10.5.0)
|
13
|
+
rspec (3.7.0)
|
14
|
+
rspec-core (~> 3.7.0)
|
15
|
+
rspec-expectations (~> 3.7.0)
|
16
|
+
rspec-mocks (~> 3.7.0)
|
17
|
+
rspec-core (3.7.1)
|
18
|
+
rspec-support (~> 3.7.0)
|
19
|
+
rspec-expectations (3.7.0)
|
20
|
+
diff-lcs (>= 1.2.0, < 2.0)
|
21
|
+
rspec-support (~> 3.7.0)
|
22
|
+
rspec-mocks (3.7.0)
|
23
|
+
diff-lcs (>= 1.2.0, < 2.0)
|
24
|
+
rspec-support (~> 3.7.0)
|
25
|
+
rspec-support (3.7.1)
|
26
|
+
rya (0.1.3)
|
27
|
+
abort_if (~> 0.2.0)
|
28
|
+
systemu (~> 2.6, >= 2.6.5)
|
29
|
+
systemu (2.6.5)
|
30
|
+
|
31
|
+
PLATFORMS
|
32
|
+
ruby
|
33
|
+
|
34
|
+
DEPENDENCIES
|
35
|
+
big_simon!
|
36
|
+
bundler (~> 1.16)
|
37
|
+
rake (~> 10.0)
|
38
|
+
rspec (~> 3.0)
|
39
|
+
|
40
|
+
BUNDLED WITH
|
41
|
+
1.16.3
|
data/README.md
ADDED
@@ -0,0 +1,37 @@
|
|
1
|
+
# BigSimon
|
2
|
+
|
3
|
+
A pipeline for finding hosts of viruses!
|
4
|
+
|
5
|
+
## Installation
|
6
|
+
|
7
|
+
Add this line to your application's Gemfile:
|
8
|
+
|
9
|
+
```ruby
|
10
|
+
gem 'big_simon'
|
11
|
+
```
|
12
|
+
|
13
|
+
And then execute:
|
14
|
+
|
15
|
+
$ bundle
|
16
|
+
|
17
|
+
Or install it yourself as:
|
18
|
+
|
19
|
+
$ gem install big_simon
|
20
|
+
|
21
|
+
## Usage
|
22
|
+
|
23
|
+
TODO
|
24
|
+
|
25
|
+
## Development
|
26
|
+
|
27
|
+
After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake spec` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.
|
28
|
+
|
29
|
+
To install this gem onto your local machine, run `bundle exec rake install`. To release a new version, update the version number in `version.rb`, and then run `bundle exec rake release`, which will create a git tag for the version, push git commits and tags, and push the `.gem` file to [rubygems.org](https://rubygems.org).
|
30
|
+
|
31
|
+
## Contributing
|
32
|
+
|
33
|
+
Bug reports and pull requests are welcome on GitHub at https://github.com/mooreryan/big_simon. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the [Contributor Covenant](http://contributor-covenant.org) code of conduct.
|
34
|
+
|
35
|
+
## Code of Conduct
|
36
|
+
|
37
|
+
Everyone interacting in the BigSimon project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the [code of conduct](https://github.com/[USERNAME]/big_simon/blob/master/CODE_OF_CONDUCT.md).
|
data/Rakefile
ADDED
data/big_simon.gemspec
ADDED
@@ -0,0 +1,30 @@
|
|
1
|
+
|
2
|
+
lib = File.expand_path("../lib", __FILE__)
|
3
|
+
$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
|
4
|
+
require "big_simon/version"
|
5
|
+
|
6
|
+
Gem::Specification.new do |spec|
|
7
|
+
spec.name = "big_simon"
|
8
|
+
spec.version = BigSimon::VERSION
|
9
|
+
spec.authors = ["Ryan Moore"]
|
10
|
+
spec.email = ["moorer@udel.edu"]
|
11
|
+
|
12
|
+
spec.summary = %q{Viral host discovery pipeline.}
|
13
|
+
spec.description = %q{Viral host discovery pipeline.}
|
14
|
+
spec.homepage = "https://github.com/mooreryan/big_simon"
|
15
|
+
|
16
|
+
# Specify which files should be added to the gem when it is released.
|
17
|
+
# The `git ls-files -z` loads the files in the RubyGem that have been added into git.
|
18
|
+
spec.files = Dir.chdir(File.expand_path('..', __FILE__)) do
|
19
|
+
`git ls-files -z`.split("\x0").reject { |f| f.match(%r{^(test|spec|features)/}) }
|
20
|
+
end
|
21
|
+
spec.bindir = "exe"
|
22
|
+
spec.executables = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
|
23
|
+
spec.require_paths = ["lib"]
|
24
|
+
|
25
|
+
spec.add_development_dependency "bundler", "~> 1.16"
|
26
|
+
spec.add_development_dependency "rake", "~> 10.0"
|
27
|
+
spec.add_development_dependency "rspec", "~> 3.0"
|
28
|
+
|
29
|
+
spec.add_runtime_dependency "rya", "~> 0.1.3"
|
30
|
+
end
|
data/bin/console
ADDED
@@ -0,0 +1,14 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require "bundler/setup"
|
4
|
+
require "big_simon"
|
5
|
+
|
6
|
+
# You can add fixtures and/or initialization code here to make experimenting
|
7
|
+
# with your gem easier. You can also use a different console, if you like.
|
8
|
+
|
9
|
+
# (If you use this, don't forget to add pry to your Gemfile!)
|
10
|
+
# require "pry"
|
11
|
+
# Pry.start
|
12
|
+
|
13
|
+
require "irb"
|
14
|
+
IRB.start(__FILE__)
|
data/bin/setup
ADDED
data/lib/big_simon.rb
ADDED
@@ -0,0 +1,100 @@
|
|
1
|
+
require "rya"
|
2
|
+
|
3
|
+
require "big_simon/version"
|
4
|
+
|
5
|
+
Time.extend Rya::CoreExtensions::Time
|
6
|
+
Process.extend Rya::CoreExtensions::Process
|
7
|
+
|
8
|
+
module BigSimon
|
9
|
+
|
10
|
+
# Project directories
|
11
|
+
ROOT = File.join __dir__, ".."
|
12
|
+
BIN = File.join ROOT, "vendor", "bin", "mac"
|
13
|
+
SPEC = File.join ROOT, "spec"
|
14
|
+
TEST_FILES = File.join SPEC, "test_files"
|
15
|
+
|
16
|
+
class Parsers
|
17
|
+
|
18
|
+
def self.vir_host_matcher fname
|
19
|
+
hosts = nil
|
20
|
+
|
21
|
+
host_info = {}
|
22
|
+
File.open(fname, "rt").each_line.with_index do |line, idx|
|
23
|
+
line.chomp!
|
24
|
+
line.sub! /,$/, "" # git rid of trailing commas
|
25
|
+
|
26
|
+
if idx.zero?
|
27
|
+
stat, *hosts = line.split ","
|
28
|
+
else
|
29
|
+
ary = line.split ","
|
30
|
+
virus = ary.shift
|
31
|
+
|
32
|
+
dists = ary.map.
|
33
|
+
with_index { |dist, idx| [hosts[idx], dist.to_f] }.
|
34
|
+
sort_by { |_, dist| dist }
|
35
|
+
|
36
|
+
best_host = dists[0][0]
|
37
|
+
|
38
|
+
host_info[virus] = {
|
39
|
+
best: best_host,
|
40
|
+
all: dists
|
41
|
+
}
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
45
|
+
host_info
|
46
|
+
end
|
47
|
+
end
|
48
|
+
|
49
|
+
class Runners
|
50
|
+
|
51
|
+
# Runs the WIsH program
|
52
|
+
#
|
53
|
+
# @raise [AbortIf::Exit] if commands fail
|
54
|
+
def self.wish exe, vir_dir, host_dir, outdir, threads
|
55
|
+
model_dir = File.join outdir, "model"
|
56
|
+
|
57
|
+
FileUtils.mkdir_p model_dir
|
58
|
+
|
59
|
+
build_model = "#{exe} " \
|
60
|
+
"-t #{threads} " \
|
61
|
+
"-c build " \
|
62
|
+
"-g #{host_dir} " \
|
63
|
+
"-m #{model_dir}"
|
64
|
+
|
65
|
+
predict = "#{exe} " \
|
66
|
+
"-t #{threads} " \
|
67
|
+
"-c predict " \
|
68
|
+
"-g #{vir_dir} " \
|
69
|
+
"-m #{model_dir} " \
|
70
|
+
"-r #{outdir} -b"
|
71
|
+
|
72
|
+
Process.run_and_time_it! "Building model", build_model
|
73
|
+
Process.run_and_time_it! "Predicting host", predict
|
74
|
+
|
75
|
+
FileUtils.rm_r model_dir if Dir.exist? model_dir
|
76
|
+
end
|
77
|
+
|
78
|
+
def self.vir_host_matcher exe, vir_dir, host_dir, outdir
|
79
|
+
FileUtils.mkdir_p outdir
|
80
|
+
|
81
|
+
cmd = "python #{exe} " \
|
82
|
+
"-v #{vir_dir} " \
|
83
|
+
"-b #{host_dir} " \
|
84
|
+
"-o #{outdir} " \
|
85
|
+
"-d 1" # only compute d2star dissimilarity
|
86
|
+
|
87
|
+
Process.run_and_time_it! "Computing d2star dissimilarity", cmd
|
88
|
+
|
89
|
+
tmp_dir = File.join outdir, "tmp"
|
90
|
+
FileUtils.rm_r tmp_dir if Dir.exist? tmp_dir
|
91
|
+
|
92
|
+
bad_files = %w[d2star_k6_main.html hostTaxa.txt_new.txt]
|
93
|
+
bad_files.each do |fname|
|
94
|
+
path = File.join outdir, fname
|
95
|
+
|
96
|
+
FileUtils.rm path if File.exist? path
|
97
|
+
end
|
98
|
+
end
|
99
|
+
end
|
100
|
+
end
|
data/vendor/bin/mac/WIsH
ADDED
Binary file
|
Binary file
|
Binary file
|
Binary file
|
@@ -0,0 +1,381 @@
|
|
1
|
+
#!/usr/bin/env python
|
2
|
+
import os,sys
|
3
|
+
import optparse
|
4
|
+
import subprocess
|
5
|
+
from subprocess import call
|
6
|
+
import time
|
7
|
+
import platform
|
8
|
+
import numpy
|
9
|
+
|
10
|
+
|
11
|
+
prog_base = os.path.split(sys.argv[0])[1]
|
12
|
+
|
13
|
+
parser = optparse.OptionParser()
|
14
|
+
parser.add_option("-v", "--virusFaDir", action = "store", type = "string", dest = "virusFaDir",
|
15
|
+
help = "the directory to the folder containing bacteria virus fasta files")
|
16
|
+
parser.add_option("-b", "--hostFaDir", action = "store", type = "string", dest = "hostFaDir",
|
17
|
+
help = "the directory to the folder containing bacteria host fasta files")
|
18
|
+
parser.add_option("-o", "--out", action = "store", type = "string", dest = "outDir",
|
19
|
+
default='./', help = "output directory")
|
20
|
+
parser.add_option("-t", "--taxa", action = "store", type = "string", dest = "hostTaxaFile",
|
21
|
+
help = "the host taxa file (including the path) ")
|
22
|
+
parser.add_option("-d", "--d2star", action = "store", type = "string", dest = "onlyD2star",
|
23
|
+
default=0, help = "compute only d2star dissimilarity? 1 for yes, 0 for no")
|
24
|
+
#parser.add_option("-u", "--continue", action = "store", type = "string", dest = "onlyComputeMeasure",
|
25
|
+
# default=0, help = "kmer count is ready only compute measures? 1 for yes, 0 for no")
|
26
|
+
#parser.add_option("-k", "--kLen", action = "store", type = "string", dest = "kLen",
|
27
|
+
# help = "the length of k-tuple")
|
28
|
+
|
29
|
+
(options, args) = parser.parse_args()
|
30
|
+
if (options.virusFaDir is None or
|
31
|
+
options.hostFaDir is None) :
|
32
|
+
sys.stderr.write(prog_base + ": ERROR: missing required command-line argument")
|
33
|
+
filelog.write(prog_base + ": ERROR: missing required command-line argument")
|
34
|
+
parser.print_help()
|
35
|
+
sys.exit(0)
|
36
|
+
|
37
|
+
|
38
|
+
## tmp file directory
|
39
|
+
if not os.path.exists(options.outDir) :
|
40
|
+
os.makedirs(options.outDir)
|
41
|
+
tmpDir = os.path.join(options.outDir, "tmp")
|
42
|
+
if not os.path.exists(tmpDir) :
|
43
|
+
os.makedirs(tmpDir)
|
44
|
+
|
45
|
+
## log file ##
|
46
|
+
filelog = open(os.path.join(tmpDir, 'vhm.log'), 'w')
|
47
|
+
|
48
|
+
## name length ##
|
49
|
+
nameLen = 93 - len(options.outDir)
|
50
|
+
#### possibly because of the kmercount folder name for each contig is too long?
|
51
|
+
|
52
|
+
|
53
|
+
#################### 0: preparation ############################
|
54
|
+
|
55
|
+
## path to the programs
|
56
|
+
vhmPath = os.path.dirname(sys.argv[0])
|
57
|
+
if len(vhmPath) == 0 :
|
58
|
+
vhmPath="./"
|
59
|
+
#print vhmPath
|
60
|
+
|
61
|
+
## kmer length and MC order
|
62
|
+
kmax = 6
|
63
|
+
order = 2
|
64
|
+
|
65
|
+
## compile c++ code if not
|
66
|
+
optSys = platform.system()
|
67
|
+
if optSys == 'Linux' :
|
68
|
+
exePath = os.path.join(vhmPath, "bin", "linux64")
|
69
|
+
elif optSys == 'Darwin' :
|
70
|
+
exePath = os.path.join(vhmPath, "bin", "macDarwin")
|
71
|
+
elif optSys == 'Windows' :
|
72
|
+
exePath = os.path.join(vhmPath, "bin", "windows64")
|
73
|
+
else :
|
74
|
+
#sys.stderr.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
75
|
+
exePath = "unknown"
|
76
|
+
|
77
|
+
|
78
|
+
|
79
|
+
## countKmer c++ code
|
80
|
+
countKmerCpp = os.path.join(vhmPath, "countKmer.cpp")
|
81
|
+
countKmerOut = os.path.join(vhmPath, "countKmer.out")
|
82
|
+
if not os.path.exists(countKmerOut) :
|
83
|
+
## recognize OS, copy exe files to vhmPath
|
84
|
+
sys.stderr.write("WARNING: can't find the file " + countKmerOut + ", try to copy one from bin \n")
|
85
|
+
filelog.write("WARNING: can't find the file " + countKmerOut + ", try to copy one from bin \n")
|
86
|
+
preCountKmerOut = os.path.join(exePath, "countKmer.out")
|
87
|
+
if exePath != "unknown" :
|
88
|
+
if os.path.exists(preCountKmerOut) :
|
89
|
+
os.system("cp " + preCountKmerOut + " " + vhmPath)
|
90
|
+
else :
|
91
|
+
sys.stderr.write( "ERROR: can't find file " + preCountKmerOut + " \n Please run make under the main directory! \n" )
|
92
|
+
filelog.write( "ERROR: can't find file " + preCountKmerOut + " \n Please run make under the main directory! \n" )
|
93
|
+
sys.exit(0)
|
94
|
+
else :
|
95
|
+
## can't recognize OS, try to compile
|
96
|
+
sys.stderr.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
97
|
+
sys.stderr.write( "Trying to compile..." + "\n")
|
98
|
+
filelog.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
99
|
+
filelog.write( "Trying to compile..." + "\n")
|
100
|
+
if os.path.exists(countKmerCpp) :
|
101
|
+
os.system("g++ " + countKmerCpp + " -o " + countKmerOut + " -std=c++0x" )
|
102
|
+
else :
|
103
|
+
sys.stderr.write( "ERROR: can't find file " + countKmerCpp + " \n Please run make under the main directory! \n" )
|
104
|
+
filelog.write( "ERROR: can't find file " + countKmerCpp + " \n Please run make under the main directory! \n" )
|
105
|
+
sys.exit(0)
|
106
|
+
|
107
|
+
os.system("chmod 777 " + countKmerOut)
|
108
|
+
filelog.flush()
|
109
|
+
|
110
|
+
## computeMeasure c++ code
|
111
|
+
computeMeasureCpp = os.path.join(vhmPath, "computeMeasure.cpp")
|
112
|
+
computeMeasureOut = os.path.join(vhmPath, "computeMeasure.out")
|
113
|
+
if not os.path.exists(computeMeasureOut) :
|
114
|
+
## recognize OS, copy exe files to vhmPath
|
115
|
+
sys.stderr.write("WARNING: can't find the file " + computeMeasureOut + ", try to copy one from bin \n")
|
116
|
+
filelog.write("WARNING: can't find the file " + computeMeasureOut + ", try to copy one from bin \n")
|
117
|
+
preComputeMeasureOut = os.path.join(exePath, "computeMeasure.out")
|
118
|
+
if exePath != "unknown" :
|
119
|
+
if os.path.exists(preComputeMeasureOut) :
|
120
|
+
os.system("cp " + preComputeMeasureOut + " " + vhmPath)
|
121
|
+
else :
|
122
|
+
sys.stderr.write( "ERROR: can't find file " + preComputeMeasureOut + " \n Please run make under the main directory! \n" )
|
123
|
+
filelog.write( "ERROR: can't find file " + preComputeMeasureOut + " \n Please run make under the main directory! \n" )
|
124
|
+
sys.exit(0)
|
125
|
+
else :
|
126
|
+
## can't recognize OS, try to compile
|
127
|
+
sys.stderr.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
128
|
+
sys.stderr.write( "Trying to compile..." + "\n")
|
129
|
+
filelog.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
130
|
+
filelog.write( "Trying to compile..." + "\n")
|
131
|
+
if os.path.exists(computeMeasureCpp) :
|
132
|
+
os.system("g++ " + computeMeasureCpp + " -o " + computeMeasureOut + " -std=c++0x" )
|
133
|
+
else :
|
134
|
+
sys.stderr.write( "ERROR: can't find file " + computeMeasureCpp + " \n Please run make under the main directory! \n" )
|
135
|
+
filelog.write( "ERROR: can't find file " + computeMeasureCpp + " \n Please run make under the main directory! \n" )
|
136
|
+
sys.exit(0)
|
137
|
+
|
138
|
+
os.system("chmod 777 " + computeMeasureOut)
|
139
|
+
filelog.flush()
|
140
|
+
|
141
|
+
## computeMeasure c++ code
|
142
|
+
computed2starCpp = os.path.join(vhmPath, "computeMeasure_onlyd2star.cpp")
|
143
|
+
computed2starOut = os.path.join(vhmPath, "computeMeasure_onlyd2star.out")
|
144
|
+
if not os.path.exists(computed2starOut) :
|
145
|
+
## recognize OS, copy exe files to vhmPath
|
146
|
+
sys.stderr.write("WARNING: can't find the file " + computed2starOut + ", try to copy one from bin \n")
|
147
|
+
preComputed2starOut = os.path.join(exePath, "computeMeasure_onlyd2star.out")
|
148
|
+
if exePath != "unknown" :
|
149
|
+
if os.path.exists(preComputed2starOut) :
|
150
|
+
os.system("cp " + preComputed2starOut + " " + vhmPath)
|
151
|
+
else :
|
152
|
+
sys.stderr.write( "ERROR: can't find file " + preComputed2starOut + " \n Please run make under the main directory! \n" )
|
153
|
+
filelog.write( "ERROR: can't find file " + preComputed2starOut + " \n Please run make under the main directory! \n" )
|
154
|
+
sys.exit(0)
|
155
|
+
else :
|
156
|
+
## can't recognize OS, try to compile
|
157
|
+
sys.stderr.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
158
|
+
sys.stderr.write( "Trying to compile..." + "\n")
|
159
|
+
filelog.write("WARNING: can't recognize the operating system" + optSys + " \n")
|
160
|
+
filelog.write( "Trying to compile..." + "\n")
|
161
|
+
if os.path.exists(computed2starCpp) :
|
162
|
+
os.system("g++ " + computed2starCpp + " -o " + computed2starOut + " -std=c++0x" )
|
163
|
+
else :
|
164
|
+
sys.stderr.write( "ERROR: can't find file " + computed2starCpp + " \n Please run make under the main directory! \n" )
|
165
|
+
filelog.write( "ERROR: can't find file " + computed2starCpp + " \n Please run make under the main directory! \n" )
|
166
|
+
sys.exit(0)
|
167
|
+
|
168
|
+
os.system("chmod 777 " + computed2starOut)
|
169
|
+
filelog.flush()
|
170
|
+
|
171
|
+
## if only compute d2star
|
172
|
+
if int(options.onlyD2star) == 1 :
|
173
|
+
computeMeasureOut=computed2starOut
|
174
|
+
else :
|
175
|
+
computeMeasureOut=computeMeasureOut
|
176
|
+
|
177
|
+
|
178
|
+
## kmer count directory
|
179
|
+
kmerCountPath = os.path.join(tmpDir, "KC")
|
180
|
+
if not os.path.exists(kmerCountPath) :
|
181
|
+
os.makedirs(kmerCountPath)
|
182
|
+
|
183
|
+
## virusFaList, hostFaList
|
184
|
+
virusFaList = os.listdir(options.virusFaDir)
|
185
|
+
#virusFaList = os.path.join(options.virusFaDir, os.listdir(options.virusFaDir))
|
186
|
+
hostFaList = os.listdir(options.hostFaDir)
|
187
|
+
|
188
|
+
## virus list file, host list file
|
189
|
+
virusListFile = os.path.join(tmpDir, "virusList")
|
190
|
+
hostListFile = os.path.join(tmpDir, "hostList")
|
191
|
+
|
192
|
+
virusListFileWrite = open(virusListFile, 'w') ## make file blank
|
193
|
+
virusListFileWrite.close()
|
194
|
+
virusListFileWrite = open(virusListFile, 'a')
|
195
|
+
|
196
|
+
hostListFileWrite = open(hostListFile, 'w') ## make file blank
|
197
|
+
hostListFileWrite.close()
|
198
|
+
hostListFileWrite = open(hostListFile, 'a')
|
199
|
+
|
200
|
+
#time.sleep(6) # delays for 10 seconds
|
201
|
+
|
202
|
+
################ 00: hostTaxa issues: ##########################
|
203
|
+
############## 1. hostName=hostFileName(with extension) ######
|
204
|
+
############## 2. hostTaxa should be no missing ########
|
205
|
+
|
206
|
+
#################### 00: if hostTaxa missing ###################
|
207
|
+
if options.hostTaxaFile is None :
|
208
|
+
hostTaxaFile = os.path.join(options.outDir, "hostTaxa.txt_new.txt")
|
209
|
+
sys.stdout.write("WARNING: no hostTaxa file provided, creating a dummy one \n")
|
210
|
+
filelog.write("WARNING: no hostTaxa file provided, creating a dummy one \n")
|
211
|
+
hostTaxaFileWrite = open(hostTaxaFile, 'w') ## make file blank
|
212
|
+
hostTaxaFileWrite.close()
|
213
|
+
hostTaxaFileWrite = open(hostTaxaFile, 'a')
|
214
|
+
|
215
|
+
hostTaxaFileWrite.write("hostNCBIName hostName hostSuperkingdom hostPhylum hostClass hostOrder hostFamily hostGenus hostSpecies\n")
|
216
|
+
for currentFileName in hostFaList :
|
217
|
+
if currentFileName.startswith('.') :
|
218
|
+
continue
|
219
|
+
if len(currentFileName) > nameLen :
|
220
|
+
sys.stdout.write( "WARNING: the file name has more than " + str(nameLen) + " letters! Use the first " + str(nameLen) + " letters as the name \n")
|
221
|
+
filelog.write( "WARNING: the file name has more than " + str(nameLen) + " letters! Use the first " + str(nameLen) + " letters as the name \n")
|
222
|
+
currentFileNameS = currentFileName[:nameLen]
|
223
|
+
else :
|
224
|
+
currentFileNameS = currentFileName
|
225
|
+
hostTaxaFileWrite.write(currentFileNameS)
|
226
|
+
for i in range(1,9) :
|
227
|
+
hostTaxaFileWrite.write("\t" + "unknown")
|
228
|
+
hostTaxaFileWrite.write("\n")
|
229
|
+
hostTaxaFileWrite.close()
|
230
|
+
options.hostTaxaFile = hostTaxaFile
|
231
|
+
|
232
|
+
################### REFORMAT hostTaxa (fill missing) ###############
|
233
|
+
else :
|
234
|
+
hostTaxaFile = os.path.join(options.outDir, os.path.basename(options.hostTaxaFile)+"_new.txt")
|
235
|
+
hostTaxaTable = numpy.genfromtxt(options.hostTaxaFile,delimiter="\t", dtype=str)
|
236
|
+
hostTaxaTable[hostTaxaTable=='']='unknown'
|
237
|
+
numpy.savetxt(hostTaxaFile, hostTaxaTable, fmt="%s", delimiter='\t', newline='\n')
|
238
|
+
|
239
|
+
filelog.flush()
|
240
|
+
|
241
|
+
#################### 1: count kmer and prepare list files ############################
|
242
|
+
#sys.stdout.write("Step 1: counting kmers \n")
|
243
|
+
|
244
|
+
start_time = time.time()
|
245
|
+
count = 0
|
246
|
+
|
247
|
+
for currentFileName in virusFaList :
|
248
|
+
if currentFileName.startswith('.') :
|
249
|
+
continue
|
250
|
+
if os.path.isdir(os.path.join(options.virusFaDir, currentFileName)) :
|
251
|
+
sys.stderr.write( "ERROR: zero bytes of file " + currentFileName + "\n")
|
252
|
+
filelog.write( "ERROR: zero bytes of file " + currentFileName + "\n")
|
253
|
+
sys.exit(0)
|
254
|
+
if len(currentFileName) > nameLen :
|
255
|
+
currentFileNameS = currentFileName[:nameLen]
|
256
|
+
else :
|
257
|
+
currentFileNameS = currentFileName
|
258
|
+
sys.stdout.write("Step 1: counting kmers for virus " + currentFileNameS + "\n")
|
259
|
+
filelog.write("Step 1: counting kmers for virus " + currentFileNameS + "\n")
|
260
|
+
for w in range(1, (kmax+1)) :
|
261
|
+
currentFilePath = os.path.join(options.virusFaDir, currentFileName)
|
262
|
+
|
263
|
+
currentKmerCountPath = os.path.join(kmerCountPath, currentFileNameS)
|
264
|
+
cmdKmer = countKmerOut + " -l -k " + str(w) + \
|
265
|
+
" -i " + currentFilePath +\
|
266
|
+
" -o " + currentKmerCountPath +\
|
267
|
+
" -s " + currentFileNameS
|
268
|
+
cmdKmerOut = subprocess.Popen(cmdKmer, shell=True, \
|
269
|
+
stderr = subprocess.PIPE, \
|
270
|
+
stdout = subprocess.PIPE)
|
271
|
+
cmdKmerOut.wait()
|
272
|
+
|
273
|
+
if len(os.listdir(currentKmerCountPath)) == ( kmax + 1 ):
|
274
|
+
virusListFileWrite.write(currentFileNameS + " " + \
|
275
|
+
currentKmerCountPath + " " +\
|
276
|
+
str(2) + "\n")
|
277
|
+
else :
|
278
|
+
sys.stderr.write( "ERROR in counting kmers for " + currentFileNameS + "\n")
|
279
|
+
filelog.write( "ERROR in counting kmers for " + currentFileNameS + "\n")
|
280
|
+
sys.exit(0)
|
281
|
+
|
282
|
+
filelog.flush()
|
283
|
+
|
284
|
+
end_time = time.time()
|
285
|
+
count += 1
|
286
|
+
#sys.stdout.write(str(end_time - start_time) + "s for " + str(count) + " seqs \n")
|
287
|
+
sys.stdout.write(" (Average time for counting kmers for one virus: " + str(round((end_time - start_time)/count,4)) + "s) \n" )
|
288
|
+
sys.stdout.write(" (ETR for counting kmers for viruses: " + str(round((end_time - start_time)/count*(len(virusFaList)-count), 4)) + "s) \n")
|
289
|
+
filelog.write(" (Average time for counting kmers for one virus: " + str(round((end_time - start_time)/count,4)) + "s) \n" )
|
290
|
+
filelog.write(" (ETR for counting kmers for viruses: " + str(round((end_time - start_time)/count*(len(virusFaList)-count), 4)) + "s) \n")
|
291
|
+
|
292
|
+
virusListFileWrite.close()
|
293
|
+
|
294
|
+
|
295
|
+
start_time = time.time()
|
296
|
+
count = 0
|
297
|
+
|
298
|
+
for currentFileName in hostFaList :
|
299
|
+
if currentFileName.startswith('.') :
|
300
|
+
continue
|
301
|
+
if os.path.isdir(os.path.join(options.hostFaDir, currentFileName)) :
|
302
|
+
sys.stderr.write( "ERROR: zero bytes of file " + currentFileName + "\n")
|
303
|
+
filelog.write( "ERROR: zero bytes of file " + currentFileName + "\n")
|
304
|
+
sys.exit(0)
|
305
|
+
if len(currentFileName) > nameLen :
|
306
|
+
currentFileNameS = currentFileName[:nameLen]
|
307
|
+
else :
|
308
|
+
currentFileNameS = currentFileName
|
309
|
+
sys.stdout.write("Step 1: counting kmers for host " + currentFileNameS + "\n")
|
310
|
+
filelog.write("Step 1: counting kmers for host " + currentFileNameS + "\n")
|
311
|
+
for w in range(1, (kmax+1)) :
|
312
|
+
currentFilePath = os.path.join(options.hostFaDir, currentFileName)
|
313
|
+
currentKmerCountPath = os.path.join(kmerCountPath, currentFileNameS)
|
314
|
+
cmdKmer = countKmerOut + " -l -k " + str(w) + \
|
315
|
+
" -i " + currentFilePath +\
|
316
|
+
" -o " + currentKmerCountPath +\
|
317
|
+
" -s " + currentFileNameS
|
318
|
+
#print cmdKmer
|
319
|
+
cmdKmerOut = subprocess.Popen(cmdKmer, shell=True, \
|
320
|
+
stderr = subprocess.PIPE, \
|
321
|
+
stdout = subprocess.PIPE)
|
322
|
+
cmdKmerOut.wait()
|
323
|
+
|
324
|
+
if len(os.listdir(currentKmerCountPath)) == ( kmax + 1 ) :
|
325
|
+
hostListFileWrite.write(currentFileNameS + " " + \
|
326
|
+
currentKmerCountPath + " " +\
|
327
|
+
str(2) + "\n")
|
328
|
+
else :
|
329
|
+
sys.stderr.write( "ERROR in counting kmers for " + currentFileNameS + "\n")
|
330
|
+
filelog.write( "ERROR in counting kmers for " + currentFileNameS + "\n")
|
331
|
+
sys.exit(0)
|
332
|
+
|
333
|
+
filelog.flush()
|
334
|
+
|
335
|
+
end_time = time.time()
|
336
|
+
count += 1
|
337
|
+
#sys.stdout.write(str(end_time - start_time) + "s for " + str(count) + " seqs \n")
|
338
|
+
sys.stdout.write(" (Average time for counting kmers for one host: " + str(round((end_time - start_time)/count, 4)) + "s) \n")
|
339
|
+
sys.stdout.write(" (ETR for counting kmers for hosts: " + str(round((end_time - start_time)/count*(len(hostFaList)-count), 4)) + "s) \n")
|
340
|
+
filelog.write(" (Average time for counting kmers for one host: " + str(round((end_time - start_time)/count, 4)) + "s) \n")
|
341
|
+
filelog.write(" (ETR for counting kmers for hosts: " + str(round((end_time - start_time)/count*(len(hostFaList)-count), 4)) + "s) \n")
|
342
|
+
|
343
|
+
hostListFileWrite.close()
|
344
|
+
|
345
|
+
|
346
|
+
|
347
|
+
#time.sleep(6) # delays for 10 seconds
|
348
|
+
|
349
|
+
################### 2: compute measures #####################
|
350
|
+
start_time = time.time()
|
351
|
+
count = 0
|
352
|
+
|
353
|
+
sys.stdout.write("Step 2: compute distance/dissimilarity measures \n")
|
354
|
+
filelog.write("Step 2: compute distance/dissimilarity measures \n")
|
355
|
+
cmdCptMeasure = computeMeasureOut + " -k " + str(kmax) + \
|
356
|
+
" -i " + virusListFile + " -j " + hostListFile + \
|
357
|
+
" -o " + options.outDir + " -t " + hostTaxaFile
|
358
|
+
print(cmdCptMeasure)
|
359
|
+
|
360
|
+
#with open(os.path.join(tmpDir, 'computeMeasureOut.log'), 'w') as filelog:
|
361
|
+
cmdCptMeasureOut = subprocess.Popen(cmdCptMeasure, shell=True, \
|
362
|
+
stderr = subprocess.PIPE, \
|
363
|
+
stdout = subprocess.PIPE)
|
364
|
+
#for c in iter(lambda: cmdCptMeasureOut.stderr.read(1), ''):
|
365
|
+
for c in iter(cmdCptMeasureOut.stderr.readline, b''):
|
366
|
+
sys.stdout.write(c.decode("utf-8"))
|
367
|
+
filelog.write(c.decode("utf-8"))
|
368
|
+
filelog.flush()
|
369
|
+
|
370
|
+
end_time = time.time()
|
371
|
+
count += 1
|
372
|
+
sys.stdout.write(" (Average time for computing dissimilarities for one virus-host pair: " + str(round((end_time - start_time)/count/len(virusFaList), 4)) + "s) \n")
|
373
|
+
sys.stdout.write(" (ETR for computing dissimilarities for virus-host pairs: " + str(round((end_time - start_time)/count*(len(hostFaList)-count), 4)) + "s) \n")
|
374
|
+
filelog.write(" (Average time for computing dissimilarities for one virus-host pair: " + str(round((end_time - start_time)/count/len(virusFaList), 4)) + "s) \n")
|
375
|
+
filelog.write(" (ETR for computing dissimilarities for virus-host pairs: " + str(round((end_time - start_time)/count*(len(hostFaList)-count), 4)) + "s) \n")
|
376
|
+
|
377
|
+
cmdCptMeasureOut.wait()
|
378
|
+
|
379
|
+
filelog.close()
|
380
|
+
|
381
|
+
sys.stdout.write("done \n")
|