aai 0.3.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 27ae62f0f9faca6746e563c76f8ca78376cd7e06
-  data.tar.gz: 0766cf35cf6654147d7a07261f2e4c9cef610062
+  metadata.gz: b2cbffaec2af6b021831515c48b227eeb9115b95
+  data.tar.gz: 537374f723948c2b42d118e3bd2cfc702b0f39ad
 SHA512:
-  metadata.gz: 4ab936b2f2a91c5b1415e2adf552a809b65487e0d6b3b611cc6551b28686109e6d402e9452cb1e6229b3c78fa16356837e1013bde573954db0f77ecfd53027e3
-  data.tar.gz: ed52d87efbcdb2aed8bd9ecb7f8da62cceb029e44a4e4f1cc61d55b777dbeb6f711583f51e065a527267edbcba381cb1f511c982785bd05624532124e6ed8a8a
+  metadata.gz: 9bc4dcd6f1ad1401d96734d7e97f6bf8a614dd474e9d61eb6973f8f148c948cbbfe47fd4671a30d80eb26345b56315912a7b94f26e1b4a2dad4f2afbbead313e
+  data.tar.gz: 6ab74991df2c29db77f8b0ddfd3c72b69428c01cfd3c2f9952f4d6687503d33caa8dcc472d1349938f41cd849639afbd4d5f4d855799d2759605529a45edef14

data/CHANGELOG.md

@@ -0,0 +1,6 @@
+# Change log
+
+Only key changes are listed here.
+
+- 0.4.0: Switch to using DIAMOND instead of BLAST
+- 0.3.0: Last version that uses blast.

data/README.md

@@ -1,7 +1,5 @@
 # AAI
 
-[![Build Status](https://travis-ci.org/mooreryan/aai.svg?branch=master)](https://travis-ci.org/mooreryan/aai)
-[![Coverage Status](https://coveralls.io/repos/github/mooreryan/aai/badge.svg?branch=master)](https://coveralls.io/github/mooreryan/aai?branch=master)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
 Calculate Seanie's multi-genome (or genome bin, or metagenome sample) amino acid similarity.
@@ -10,8 +8,13 @@ Calculate Seanie's multi-genome (or genome bin, or metagenome sample) amino acid
 
 The following programs must be installed and on your `PATH` for `aai` to work.
 
-- GNU Parallel
-- NCBI Blast suite
+For versions `>= 0.4`
+
+- [DIAMOND](https://github.com/bbuchfink/diamond/)
+
+For versions `< 0.4`
+
+- [NCBI Blast suite](https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastDocs&DOC_TYPE=Download)
 
 ## Installation
 
@@ -38,13 +41,14 @@ And then execute:
 ### Example
 
 ```
-$ ruby exe/aai.rb --infiles *.fa
+$ ruby exe/aai.rb --infiles *.fa --outdir aai_output
 ```
 
 ### Options
 
 ```
 Options:
+  -c, --cpus=<i>        Number of CPUs to use (default: 1)
   -i, --infiles=<s+>    Input files
   -o, --outdir=<s>      Output directory (default: .)
   -b, --basename=<s>    Base name for output file (default: aai_scores)

data/aai.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "yard", "~> 0.9.9"
 
   spec.add_runtime_dependency "abort_if", "~> 0.2.0"
-  spec.add_runtime_dependency "parallel", "~> 1.6", ">= 1.6.1"
+  # spec.add_runtime_dependency "parallel", "~> 1.6", ">= 1.6.1"
   spec.add_runtime_dependency "parse_fasta", "~> 2.2"
   spec.add_runtime_dependency "systemu", "~> 2.6", ">= 2.6.5"
   spec.add_runtime_dependency "trollop", "~> 2.1", ">= 2.1.2"

data/lib/aai.rb

@@ -1,6 +1,5 @@
 require "abort_if"
 require "systemu"
-require "parallel"
 require "parse_fasta"
 
 require "aai/core_extensions"
@@ -21,12 +20,17 @@ module Aai
   EVALUE_CUTOFF = 1e-3
   LENGTH_CUTOFF = 70 # actually is 70 percent
 
+  # If a blast job fails, it will retry once. If it fails again, it
+  # will be ignored by the rest of the pipeline.
   def blast_permutations! fastas, blast_dbs, cpus=4
     file_permutations = one_way_combinations fastas, blast_dbs, true
     file_permutations = file_permutations.select do |f1, f2|
       genome_from_fname(f1) != genome_from_fname(f2)
     end
 
+    completed_outf_names = []
+    failed_jobs = []
+
     first_files = file_permutations.map(&:first)
     second_files = file_permutations.map(&:last)
 
@@ -49,22 +53,61 @@ module Aai
       "#{f1}____#{f2}.aai_blastp"
     end
 
-    parallel_args = first_files.length.times.map do |idx|
+    args = first_files.length.times.map do |idx|
       [first_files[idx], second_files[idx], outf_names[idx]]
     end
 
     Time.time_it "Running blast jobs" do
-      Parallel.each(parallel_args, in_processes: cpus) do |infiles|
+      args.each_with_index do |infiles, idx|
         query = infiles[0]
         db    = infiles[1]
         out   = infiles[2]
 
-        cmd = "blastp -outfmt 6 -query #{query} -db #{db} -out #{out} -evalue #{EVALUE_CUTOFF}"
-        Process.run_it! cmd
+        cmd = "diamond blastp --threads #{cpus} --outfmt 6 " +
+              "--query #{query} --db #{db} --out #{out} " +
+              "--evalue #{EVALUE_CUTOFF}"
+
+        exit_status = Process.run_it cmd
+
+        if exit_status.zero?
+          completed_outf_names << out
+        else
+          failed_jobs << idx
+          AbortIf.logger.warn { "Blast job failed. Non-zero exit status " +
+                                "(#{exit_status}) " +
+                                "when running '#{cmd}'. " +
+                                "Will retry at end." }
+        end
+      end
+    end
+
+    if failed_jobs.count > 0
+      Time.time_it "Retrying failed blast jobs" do
+        # retry failed jobs once
+        failed_jobs.each do |idx|
+          query = args[idx][0]
+          db    = args[idx][1]
+          out   = args[idx][2]
+
+          cmd = "diamond blastp --threads #{cpus} --outfmt 6 " +
+                "--query #{query} --db #{db} --out #{out} " +
+                "--evalue #{EVALUE_CUTOFF}"
+
+          exit_status = Process.run_it cmd
+
+          if exit_status.zero?
+            completed_outf_names << out
+          else
+            AbortIf.logger.error { "Retrying blast job failed. " +
+                                   "Non-zero exit status " +
+                                   "(#{exit_status}) " +
+                                   "when running '#{cmd}'." }
+          end
+        end
       end
     end
 
-    outf_names
+    completed_outf_names
   end
 
   # Make blast dbs given an array of filenames.
@@ -78,8 +121,9 @@ module Aai
     outfiles = fnames.map { |fname| fname + suffix }
 
     Time.time_it "Making blast databases" do
-      Parallel.each(fnames, in_processes: cpus) do |fname|
-        cmd = "makeblastdb -in #{fname} -out #{fname}#{BLAST_DB_SUFFIX} -dbtype prot"
+      fnames.each do |fname|
+        cmd = "diamond makedb --threads #{cpus} --in #{fname} " +
+              "--db #{fname}#{BLAST_DB_SUFFIX}"
 
         Process.run_it! cmd
       end
@@ -103,13 +147,15 @@ module Aai
       clean_fnames << clean_fname
       File.open(clean_fname, "w") do |f|
         Object::ParseFasta::SeqFile.open(fname).each_record do |rec|
-          header =
-            annotate_header clean_header(rec.header),
-                            File.basename(fname)
+          unless bad_seq? rec.seq
+            header =
+              annotate_header clean_header(rec.header),
+                              File.basename(fname)
 
-          seq_lengths[header] = rec.seq.length
+            seq_lengths[header] = rec.seq.length
 
-          f.puts ">#{header}\n#{rec.seq}"
+            f.puts ">#{header}\n#{rec.seq}"
+          end
         end
       end
     end
@@ -269,6 +315,13 @@ module Aai
 
   private
 
+  # this is to account for the weird IMG error. Some seqs will
+  # not have an actual protein, rather it will be "No sequence
+  # found"
+  def bad_seq? seq
+    seq.downcase.include? "nosequencefound"
+  end
+
   def two_way_hit? hit1, hit2
     hit1[:query_name] == hit2[:target_name] &&
       hit1[:query_genome] == hit2[:target_genome]

data/lib/aai/utils.rb

@@ -1,5 +1,9 @@
 module Aai
   module Utils
+    def clean_str str
+      str.strip.gsub(/[^\p{Alnum}_]+/, "_").gsub(/_+/, "_")
+    end
+
     # Raises SystemExit if one of the fnames does not exist.
     def check_files fnames
       fnames.each do |fname|

data/lib/aai/version.rb

@@ -1,5 +1,5 @@
 module Aai
-  VERSION = "0.3.0"
+  VERSION = "0.4.0"
   COPYRIGHT = "2017 Ryan Moore"
   CONTACT   = "moorer@udel.edu"
   WEBSITE   = "https://github.com/mooreryan/aai"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: aai
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Ryan Moore
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-06-23 00:00:00.000000000 Z
+date: 2017-06-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -94,26 +94,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: 0.2.0
-- !ruby/object:Gem::Dependency
-  name: parallel
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.6'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.6.1
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.6'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.6.1
 - !ruby/object:Gem::Dependency
   name: parse_fasta
   requirement: !ruby/object:Gem::Requirement
@@ -180,6 +160,7 @@ files:
 - ".gitignore"
 - ".rspec"
 - ".travis.yml"
+- CHANGELOG.md
 - CODE_OF_CONDUCT.md
 - Gemfile
 - LICENSE