miga-base 0.3.2.1 → 0.3.2.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 82b444c1fd986a4bc95f5cca2e2485b2fb0dd29b82d7424ef7615e41d9f901d3
-  data.tar.gz: 8bc59e799f8af6e13dc1a3e0653743490c44189e056fbe8a9f86dc53b029a6bf
+  metadata.gz: 83a243d552f8c0f850cd46e827069e2a0b6abb5b8ed3f2cda41cec21a5b93c85
+  data.tar.gz: 85d9838dc0f9708d50e7d4375b8e641bc85ef096d95c88fcc43e8b5d1421e449
 SHA512:
-  metadata.gz: a3aabd7fa5fd037b4e427b131c311a66a103f68f5ff0aaf4907abf4663ed285e7010d89a10d2443e72f4d9c1f0a032ba112faf961972da34bc44b449a3531332
-  data.tar.gz: 4b1f302f219addc604f66526f72171f711c1e89252a7f152d68e839d4235ab723a43811f16664b81dee77b91c149fd38e06e54cca7d8412bf84594f26de36dcb
+  metadata.gz: 7705f7987ae8a13a6664d8d3a9ca586bbbebc5447c5858fb5c9980c9d31d97a57e779e6987d4ca49dc06b7ccbb8602f22a2747cb1f55acfda3a5bf4edaa1798f
+  data.tar.gz: 3920ee4cdbd9b6666e0d0233c3557218ad30abe4a4e9276de8db9532da628626ba764080766bf27ca5cee335c892cf7c72d610d2fc356dd2da464b24c881c228

data/lib/miga/project/result.rb

@@ -118,8 +118,8 @@ module MiGA::Project::Result
       return nil unless result_files_exist?(base, %w[.ogs .stats])
       r = MiGA::Result.new("#{base}.json")
       r.add_file(:ogs, "miga-project.ogs")
+      r.add_file(:abc, "miga-project.abc")
       r.add_file(:stats, "miga-project.stats")
-      r.add_file(:rbm, "miga-project.rbm")
       r.add_file(:core_pan, "miga-project.core-pan.tsv")
       r.add_file(:core_pan_plot, "miga-project.core-pan.pdf")
       r

data/lib/miga/version.rb

@@ -10,7 +10,7 @@ module MiGA
   # - Float representing the major.minor version.
   # - Integer representing gem releases of the current version.
   # - Integer representing minor changes that require new version number.
-  VERSION = [0.3, 2, 1]
+  VERSION = [0.3, 2, 3]
   
   ##
   # Nickname for the current major.minor version.

data/scripts/aai_distances.bash

@@ -42,4 +42,4 @@ gzip -9 -f miga-project.txt
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/ani_distances.bash

@@ -41,4 +41,4 @@ gzip -9 -f miga-project.txt
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/assembly.bash

@@ -50,4 +50,4 @@ FastA.length.pl "$DATASET.AllContigs.fna" | awk '$2>=1000{print $1}' \
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/cds.bash

@@ -42,4 +42,4 @@ gzip -9 -f "$DATASET.gff3"
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/clade_finding.bash

@@ -44,4 +44,4 @@ fi
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/distances.bash

@@ -17,4 +17,4 @@ ruby -I "$MIGA/lib" "$MIGA/utils/distances.rb" "$PROJECT" "$DATASET"
 # Finalize
 rm -R "$TMPDIR"
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/essential_genes.bash

@@ -45,4 +45,4 @@ fi
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/haai_distances.bash

@@ -41,4 +41,4 @@ gzip -9 -f miga-project.txt
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/miga.bash

@@ -17,6 +17,8 @@ done
 [[ -n $DATASET ]] \
       && miga add -P "$PROJECT" -D "$DATASET" -m "_step=$SCRIPT" --update
 
+true
+
 #if [[ "$RUNTYPE" == "qsub" ]] ; then
 #elif [[ "$RUNTYPE" == "msub" ]] ; then
 #fi

data/scripts/mytaxa.bash

@@ -98,4 +98,4 @@ fi
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/mytaxa_scan.bash

@@ -92,4 +92,4 @@ fi
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/ogs.bash

@@ -14,25 +14,28 @@ miga date > "miga-project.start"
 DS=$(miga list_datasets -P "$PROJECT" --ref --no-multi)
 if [[ ! -s miga-project.ogs ]] ; then
   # Extract RBMs
-  [[ -d miga-project.rbm ]] || mkdir miga-project.rbm
-  echo -n "" > miga-project.log
-  for i in $DS ; do
-    for j in $DS ; do
-      file="miga-project.rbm/$i-$j.rbm"
-      [[ -s $file ]] && continue
-      echo "SELECT id1,id2,id,0,0,0,0,0,0,0,evalue,bitscore from rbm" \
-        "where seq1='$i' and seq2='$j' ;" \
-        | sqlite3 "../../09.distances/02.aai/$i.db" | tr "\\|" "\\t" \
-        > "$file"
-      [[ -s "$file" ]] || rm "$file"
+  if [[ ! -s miga-project.abc ]] ; then
+    [[ -d miga-project.tmp ]] || mkdir miga-project.tmp
+    for i in $DS ; do
+      file="miga-project.tmp/$i.abc"
+      [[ -s "$file" ]] && continue
+      echo "SELECT seq1,id1,seq2,id2,bitscore from rbm;" \
+        | sqlite3 "../../09.distances/02.aai/$i.db" | tr "\\|" " " \
+        | awk '{ print $1">"$2"'"\\t"'"$3">"$4"'"\\t"'"$5 }' \
+        > "$file.tmp"
+      mv "$file.tmp" "$file"
     done
-    echo "$i" >> miga-project.log
-  done
+    cat miga-project.tmp/*.abc > miga-project.abc
+  fi
+  rm -rf miga-project.tmp
 
   # Estimate OGs and Clean RBMs
-  ogs.mcl.rb -o miga-project.ogs -d miga-project.rbm -t "$CORES"
-  [[ $(miga about -P "$PROJECT" -m clean_ogs) == "false" ]] \
-    || rm -rf miga-project.rbm
+  ogs.mcl.rb -o miga-project.ogs --abc miga-project.abc -t "$CORES"
+  if [[ $(miga about -P "$PROJECT" -m clean_ogs) == "false" ]] ; then
+    rm miga-project.abc
+  else
+    gzip -9 miga-project.abc
+  fi
 fi
 
 # Calculate Statistics
@@ -43,4 +46,4 @@ Rscript "$MIGA/utils/core-pan-plot.R" \
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/project_stats.bash

@@ -22,4 +22,4 @@ ruby -I "$MIGA/lib" \
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/read_quality.bash

@@ -31,4 +31,4 @@ rm -f "../02.trimmed_reads/$b".[12].fastq
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/ssu.bash

@@ -33,4 +33,4 @@ fi
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/stats.bash

@@ -21,4 +21,4 @@ done
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/subclades.bash

@@ -24,4 +24,4 @@ ruby "$MIGA/utils/subclades-compile.rb" . \
 
 # Finalize
 miga date > "miga-project.done"
-miga add_result -P "$PROJECT" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -r "$SCRIPT" -f

data/scripts/taxonomy.bash

@@ -14,8 +14,9 @@ cd "$DIR"
 miga date > "$DATASET.start"
 
 # Run
-ruby "$MIGA/utils/distances.rb" "$PROJECT" "$DATASET" run_taxonomy=1
+ruby -I "$MIGA/lib" \
+      "$MIGA/utils/distances.rb" "$PROJECT" "$DATASET" run_taxonomy=1
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/trimmed_fasta.bash

@@ -47,4 +47,4 @@ done
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/scripts/trimmed_reads.bash

@@ -57,4 +57,4 @@ rm -f "$b".[12].*.discard
 
 # Finalize
 miga date > "$DATASET.done"
-miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
+miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT" -f

data/utils/distance/pipeline.rb

@@ -25,7 +25,7 @@ module MiGA::DistanceRunner::Pipeline
         end
       end
     end
-    classif = File.expand_path("miga-project.sc-#{val_cls}", classif)
+    classif = "#{classif}/miga-project.sc-#{val_cls}"
     result_fh.puts [val_cls, val_med, max_val, classif].join("\t")
     classify(clades, classif, metric, result_fh, val_cls)
   end

data/utils/distance/runner.rb

@@ -21,13 +21,14 @@ class MiGA::DistanceRunner
     @project = MiGA::Project.load(project_path) or
           raise "No project at #{project_path}"
     @dataset = project.dataset(dataset_name)
-    @home = File.expand_path("data/09.distances", project.path)
+    @home = File.expand_path('data/09.distances', project.path)
     # Default opts
-    @opts[:aai_save_rbm] ||= ENV.fetch("MIGA_AAI_SAVE_RBM") do
-      project.is_clade? ? "save-rbm" : "no-save-rbm"
+    @opts[:aai_save_rbm] ||= ENV.fetch('MIGA_AAI_SAVE_RBM') do
+      project.is_clade? ? 'save-rbm' : 'no-save-rbm'
     end
     @opts[:thr] ||= ENV.fetch("CORES"){ 2 }.to_i
     if opts[:run_taxonomy] && project.metadata[:ref_project]
+      @home = File.expand_path('05.taxonomy', @home)
       @ref_project = MiGA::Project.load(project.metadata[:ref_project])
     end
     @ref_project ||= project
@@ -85,7 +86,7 @@ class MiGA::DistanceRunner
           r = ln.chomp.split("\t")
           next unless r[1].to_i==val_cls
           target = ref_project.dataset(r[0])
-          aai = (metric==:aai) ? aai(target) : 100.0
+          aai = (v[1]==:aai) ? aai(target) : 100.0
           ani(target) if aai >= 90.0
         end
       end

data/utils/subclades.R

@@ -52,6 +52,7 @@ subclades <- function(ani_file, out_base, thr=1, ani=c()) {
   s.avg.z <- (s[1,]-mean(s[1,]))/(sd(s[1,])+0.0001)
   s.neg.z <- (s[2,]-mean(s[2,]))/(sd(s[2,])+0.01)
   ds <- s.avg.z - s.neg.z - 2/(1:length(k)) - (1:length(k))/50
+  if(mean(s[1,]<0)<0.75) ds[s[1,]<0] <- mean(ds) # <- k's with negative average
   top.n <- k[which.max(ds)]
   
   # Classify genomes

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 0.3.2.1
+  version: 0.3.2.3
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-06-28 00:00:00.000000000 Z
+date: 2018-07-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rest-client
@@ -167,9 +167,6 @@ files:
 - lib/miga/tax_index.rb
 - lib/miga/taxonomy.rb
 - lib/miga/version.rb
-- scripts/_distances_functions.bash
-- scripts/_distances_noref_nomulti.bash
-- scripts/_distances_ref_nomulti.bash
 - scripts/aai_distances.bash
 - scripts/ani_distances.bash
 - scripts/assembly.bash

data/scripts/_distances_functions.bash

@@ -1,104 +0,0 @@
-#!/bin/bash
-# Available variables: $PROJECT, $DATASET, $RUNTYPE, $MIGA, $CORES, $TMPDIR,
-#                      $NOMULTI, $REF
-
-set -e
-
-if [[ ! -n $MIGA_AAI_SAVE_RBM ]] ; then
-  MIGA_AAI_SAVE_RBM="save-rbm"
-  if [[ -n $PROJECT ]] ; then
-    if [[ $(miga project_info -P "$PROJECT" -m type) != "clade" ]] ; then
-      MIGA_AAI_SAVE_RBM="no-save-rbm"
-    fi
-  fi
-fi
-
-fx_exists miga-make_empty_aai_db || function miga-make_empty_aai_db {
-  local DB=$1
-  echo "create table if not exists aai(seq1 varchar(256), seq2 varchar(256)," \
-    " aai float, sd float, n int, omega int);" | sqlite3 "$DB"
-}
-
-fx_exists miga-ds_name || function miga-ds_name {
-  basename "$1" | perl -pe "s/[^A-Za-z0-9_].*//"
-}
-
-fx_exists miga-aai || function miga-aai {
-  local F1=$1
-  local F2=$2
-  local TH=$3
-  local DB=$4
-  local N1
-  N1=$(miga-ds_name "$F1")
-  local N2
-  N2=$(miga-ds_name "$F2")
-  aai.rb -1 "$F1" -2 "$F2" -t "$TH" -a --lookup-first -S "$DB" --name1 "$N1" \
-    --name2 "$N2" --$MIGA_AAI_SAVE_RBM || echo "0"
-}
-
-fx_exists miga-ani || function miga-ani {
-  local F1=$1
-  local F2=$2
-  local TH=$3
-  local DB=$4
-  local N1
-  N1=$(miga-ds_name "$F1")
-  local N2
-  N2=$(miga-ds_name "$F2")
-  ani.rb -1 "$F1" -2 "$F2" -t "$TH" -a --no-save-regions --no-save-rbm \
-    --lookup-first -S "$DB" --name1 "$N1" --name2 "$N2" || echo "0"
-}
-
-fx_exists miga-haai || function miga-haai {
-  local F1=$1
-  local F2=$2
-  local TH=$3
-  local DB=$4
-  local AAI_DB=$5
-  local N1
-  N1=$(miga-ds_name "$F1")
-  local N2
-  N2=$(miga-ds_name "$F2")
-  local HAAI
-  HAAI=$(MIGA_AAI_SAVE_RBM="no-save-rbm" miga-aai "$F1" "$F2" "$TH" "$DB")
-  if [[ "$HAAI" != "" && $(perl -e "print 1 if '$HAAI' <= 90") == "1" ]] ; then
-    local AAI
-    AAI=$(perl -e "print (100-exp(2.435076 + 0.4275193*log(100-$HAAI)))")
-    [[ ! -s $AAI_DB ]] && miga-make_empty_aai_db "$AAI_DB"
-    echo "insert into aai values('$N1','$N2','$AAI',0,0,0);" | sqlite3 "$AAI_DB"
-    echo "$AAI"
-  fi
-}
-
-fx_exists miga-haai_or_aai || function miga-haai_or_aai {
-  local FH1=$1
-  local FH2=$2
-  local DBH=$3
-  local  F1=$4
-  local  F2=$5
-  local  DB=$6
-  local  TH=$7
-  local AAI
-  AAI=$(miga-haai "$FH1" "$FH2" "$TH" "$DBH" "$DB")
-  [[ "${AAI%.*}" -le 0 ]] && AAI=$(miga-aai "$F1" "$F2" "$TH" "$DB")
-  echo "$AAI"
-}
-
-fx_exists miga-val_from_db || function miga-val_from_db {
-  local N1=$1
-  local N2=$2
-  local DB=$3
-  local MT=$4
-  if [[ -s $DB ]] ; then
-    echo "select $MT from $MT where seq1='$N1' and seq2='$N2';" \
-      | sqlite3 "$DB" || echo 0
-  fi
-}
-
-fx_exists miga-aai_from_db || function miga-aai_from_db {
-  miga-val_from_db "$1" "$2" "$3" aai
-}
-
-fx_exists miga-ani_from_db || function miga-ani_from_db {
-  miga-val_from_db "$1" "$2" "$3" ani
-}

data/scripts/_distances_noref_nomulti.bash

@@ -1,149 +0,0 @@
-#!/bin/bash
-# Available variables: $PROJECT, $DATASET, $RUNTYPE, $MIGA, $CORES, $TMPDIR,
-#                      $NOMULTI, $REF
-
-set -e
-
-# Deal with previous runs (if any)
-exists "$DATASET".haai.db && cp "$DATASET".haai.db "$TMPDIR"
-exists "$DATASET".a[an]i.db && cp "$DATASET".a[an]i.db "$TMPDIR"
-exists "$DATASET".a[an]i.9[05] && rm "$DATASET".a[an]i.9[05]
-N=0
-fx_exists miga-checkpoint_n || function miga-checkpoint_n {
-  let N=$N+1
-  if [[ $N -ge 10 ]] ; then
-    for metric in haai aai ani ; do
-      if [[ -s $TMPDIR/$DATASET.$metric.db ]] ; then
-        echo "select count(*) from ${metric#h};" \
-          | sqlite3 "$TMPDIR/$DATASET.$metric.db" \
-          >/dev/null || exit 1
-        cp "$TMPDIR/$DATASET.$metric.db" .
-      fi
-    done
-    N=0
-  fi
-}
-
-fx_exists miga-noref_haai_or_aai || function miga-noref_haai_or_aai {
-  local Q=$1
-  local S=$2
-  [[ -s $TMPDIR/$Q.faa ]] \
-    || cp "$PROJECT/data/06.cds/$Q.faa" "$TMPDIR/$Q.faa"
-  miga-haai_or_aai "$PROJECT/$ESS/$Q.ess.faa" "$S_PROJ/$ESS/$S.ess.faa" \
-    "$TMPDIR/$Q.haai.db" "$TMPDIR/$Q.faa" "$S_PROJ/data/06.cds/$S.faa" \
-    "$TMPDIR/$Q.aai.db" "$CORES"
-}
-
-fx_exists miga-noref_ani || function miga-noref_ani {
-  local Q=$1
-  local S=$2
-  [[ -s "$TMPDIR/$Q.LargeContigs.fna" ]] \
-    || cp "$PROJECT/data/05.assembly/$Q.LargeContigs.fna" \
-          "$TMPDIR/$Q.LargeContigs.fna"
-  miga-ani "$TMPDIR/$Q.LargeContigs.fna" \
-    "$S_PROJ/data/05.assembly/$S.LargeContigs.fna" \
-    "$CORES" "$TMPDIR/$Q.ani.db"
-}
-
-# Calculate the classification-informed AAI/ANI traverse (if not classified)
-ESS="data/07.annotation/01.function/01.essential"
-if [[ $(miga about -P "$S_PROJ" -m type) != "clade" ]] ; then
-  # Classify aai-clade (if project type is not clade)
-  CLADES="$S_PROJ/data/10.clades/01.find"
-  METRIC="aai"
-  REF_TABLE="$S_PROJ/data/09.distances/02.aai/miga-project.txt.gz"
-else
-  # Classify ani-clade (if project type is clade)
-  CLADES="$S_PROJ/data/10.clades/02.ani"
-  METRIC="ani"
-  REF_TABLE="$S_PROJ/data/09.distances/03.ani/miga-project.txt.gz"
-fi
-
-CLASSIF="."
-[[ -e "$DATASET.$METRIC-medoids.tsv" ]] && rm "$DATASET.$METRIC-medoids.tsv"
-[[ -s "$CLADES/$CLASSIF/miga-project.medoids" ]] || \
-  touch "$DATASET.$METRIC-medoids.tsv" "${DATASET}.${METRIC}.db"
-while [[ -s "$CLADES/$CLASSIF/miga-project.medoids" ]] ; do
-  MAX_VAL=0
-  VAL_MED=""
-  VAL_CLS=""
-  i_n=0
-  while read -r i ; do
-    let i_n=$i_n+1
-    if [[ $METRIC == "aai" ]] ; then
-      VAL=$(miga-noref_haai_or_aai "$DATASET" "$i")
-    else
-      VAL=$(miga-noref_ani "$DATASET" "$i")
-    fi
-    miga-checkpoint_n
-    if [[ $(perl -e "print 1 if '$VAL' >= '$MAX_VAL'") == "1" ]] ; then
-      MAX_VAL=$VAL
-      VAL_MED=$i
-      VAL_CLS=$i_n
-      echo "[$CLASSIF] New max: $VAL_MED ($VAL_CLS): $MAX_VAL"
-    fi
-  done < "$CLADES/$CLASSIF/miga-project.medoids"
-  CLASSIF="$CLASSIF/miga-project.sc-$VAL_CLS"
-  echo "$VAL_CLS	$VAL_MED	$MAX_VAL	$CLASSIF" \
-    >> "$DATASET.$METRIC-medoids.tsv"
-done
-
-# Calculate all the AAIs/ANIs against the lowest subclade (if classified)
-if [[ "$CLASSIF" != "." ]] ; then
-  PAR=$(dirname "$CLADES/$CLASSIF")/miga-project.classif
-  if [[ -s "$PAR" ]] ; then
-    while read -r i ; do
-      if [[ $METRIC == "aai" ]] ; then
-        AAI=$(miga-noref_haai_or_aai "$DATASET" "$i")
-      else
-        AAI=100
-      fi
-      if [[ $(perl -e "print 1 if '$AAI' >= 90") == "1" ]] ; then
-        miga-noref_ani "$DATASET" "$i"
-      fi
-      miga-checkpoint_n
-    done < <(awk "\$2==$VAL_CLS{print \$1}" < "$PAR")
-  fi
-fi
-
-# Finalize
-N=11
-miga-checkpoint_n
-
-# Build tree with medoids
-if [[ -s "${DATASET}.${METRIC}.db" ]] ; then
-  echo "select seq2 from $METRIC;" | sqlite3 "${DATASET}.${METRIC}.db" \
-    | sort | uniq > "${DATASET}.tmp0"
-  perl -pe "s/^/^/" "${DATASET}.tmp0" | perl -pe "s/$/\\t/" \
-    > "${DATASET}.tmp1"
-  perl -pe "s/^/\\t/" "${DATASET}.tmp0" | perl -pe "s/$/\\t/" \
-    > "${DATASET}.tmp2"
-  echo "a b value" | tr " " "\\t" > "${DATASET}.txt"
-  gzip -c -d "$REF_TABLE" | cut -f 2-4 \
-    | grep -f "${DATASET}.tmp1" | grep -f "${DATASET}.tmp2" \
-    >> "${DATASET}.txt"
-  echo "select seq1, seq2, $METRIC from $METRIC;" \
-    | sqlite3 "${DATASET}.${METRIC}.db"  | tr "\\|" "\\t" \
-    >> "${DATASET}.txt"
-  "$MIGA/utils/ref-tree.R" "${DATASET}.txt" "$DATASET" "$DATASET"
-  rm "$DATASET".tmp[012] "${DATASET}.txt"
-fi
-
-# Test taxonomy
-(
-  trap 'rm "$DATASET.json" "$DATASET.done"' EXIT
-  FLAGS=""
-  [[ "$PROJECT" == "$S_PROJ" ]] || FLAGS="--ref-project"
-  miga date > "$DATASET.done"
-  miga add_result -P "$PROJECT" -D "$DATASET" -r "$SCRIPT"
-  miga tax_test -P "$PROJECT" -D "$DATASET" -t intax \
-    $FLAGS > "$DATASET.intax.txt"
-)
-
-# Transfer taxonomy
-TAX_PVALUE=$(miga about -P "$PROJECT" -m tax_pvalue)
-[[ "$TAX_PVALUE" == "?" ]] && TAX_PVALUE="0.05"
-NEW_TAX=$(tail -n +6 "$DATASET.intax.txt" | head -n -3 \
-  | awk '$3<'$TAX_PVALUE'{print $1":"$2}' | grep -v "?" \
-  | tr "\\n" ' ' | perl -pe 's/ *$//')
-miga tax_set -P "$PROJECT" -D "$DATASET" -s "$NEW_TAX"

data/scripts/_distances_ref_nomulti.bash

@@ -1,79 +0,0 @@
-#!/bin/bash
-# Available variables: $PROJECT, $DATASET, $RUNTYPE, $MIGA, $CORES, $TMPDIR,
-# 			$NOMULTI, $REF
-
-set -e
-
-fx_exists miga-checkpoint_n || function miga-checkpoint_n {
-  if [[ $N -eq 10 ]] ; then
-    for t in 01.haai 02.aai 03.ani ; do
-      if [[ -s $TMPDIR/$t.db ]] ; then
-        tab="aai"
-        [[ "$t" == "03.ani" ]] && tab="ani"
-        echo "select count(*) from $tab;" \
-          | sqlite3 "$TMPDIR/$t.db" \
-          >/dev/null || exit 1
-        cp "$TMPDIR/$t.db" "$t/$DATASET.db"
-      fi
-    done
-    N=0
-  fi
-  let N=$N+1
-}
-
-ESS="../07.annotation/01.function/01.essential"
-
-# Initialize temporals
-for t in 01.haai 02.aai 03.ani ; do
-  [[ -s $t/$DATASET.db ]] && cp "$t/$DATASET.db" "$TMPDIR/$t.db"
-done
-N=1
-
-# Traverse "nearly-half" of the ref-datasets using first-come-first-served
-for i in $(miga list_datasets -P "$PROJECT" --ref --no-multi) ; do
-  echo "[ $(date "+%Y-%m-%d %H:%M:%S %z") ] $i"
-  AAI=""; ANI="";
-  # Check if the i-th dataset is ready
-  [[ -s $ESS/$i.done && -s $ESS/$i.json ]] || continue
-  # Check if this is done (e.g., in a previous failed iteration)
-  AAI=$(miga-aai_from_db "$DATASET" "$i" "$TMPDIR/02.aai.db")
-  # Try the other direction
-  [[ "${AAI%.*}" -le 0 ]] \
-    && AAI=$(miga-aai_from_db "$i" "$DATASET" "02.aai/$i.db")
-  # Try with hAAI
-  if [[ "${AAI%.*}" -le 0 ]] ; then
-    [[ -e "$TMPDIR/$DATASET.ess.faa" ]] \
-      || cp "$ESS/$DATASET.ess.faa" "$TMPDIR/$DATASET.ess.faa"
-    AAI=$(miga-haai "$TMPDIR/$DATASET.ess.faa" "$ESS/$i.ess.faa" \
-      "$CORES" "$TMPDIR/01.haai.db" "$TMPDIR/02.aai.db")
-  fi
-  # Try with complete AAI
-  if [[ "${AAI%.*}" -le 0 ]] ; then
-    [[ -e "$TMPDIR/$DATASET.faa" ]] \
-      || cp "../06.cds/$DATASET.faa" "$TMPDIR/$DATASET.faa"
-    AAI=$(miga-aai "$TMPDIR/$DATASET.faa" "../06.cds/$i.faa" \
-      "$CORES" "$TMPDIR/02.aai.db")
-  fi
-  # Check if ANI is meaningful
-  if [[ -e "../05.assembly/$DATASET.LargeContigs.fna" \
-      && -e "../05.assembly/$i.LargeContigs.fna" \
-      && $(perl -e "print 1 if '$AAI' >= 90") == "1" ]] ; then
-    # Check if this is done (e.g., in a previous failed iteration)
-    ANI=$(miga-ani_from_db "$DATASET" "$i" "$TMPDIR/03.ani.db")
-    # Try the other direction
-    [[ "${ANI%.*}" -le 0 ]] \
-      && ANI=$(miga-ani_from_db "$i" "$DATASET" "03.ani/$i.db")
-    # Calculate it
-    if [[ "${ANI%.*}" -le 0 ]] ; then
-      [[ -e "$TMPDIR/$DATASET.LargeContigs.fna" ]] \
-        || cp "../05.assembly/$DATASET.LargeContigs.fna" \
-          "$TMPDIR/$DATASET.LargeContigs.fna"
-      ANI=$(miga-ani "$TMPDIR/$DATASET.LargeContigs.fna" \
-        "../05.assembly/$i.LargeContigs.fna" "$CORES" "$TMPDIR/03.ani.db")
-    fi
-  fi
-  miga-checkpoint_n
-done
-N=10
-miga-checkpoint_n
-