Test BLAST+ backend transition dependencies

.github/workflows/continuous_integration.yml

@@ -25,12 +25,41 @@ jobs:
 
       - name: Install selenoprofiles4 in Conda environment
         run: |
-          conda install -n sp4 -y -c mmariotti -c anaconda -c bioconda -c biobuilds selenoprofiles4
+          conda install -n sp4 -y -c mmariotti -c anaconda -c bioconda -c biobuilds selenoprofiles4 blast
           pip install -e .[addons]
 
+      - name: Check BLAST+ backend is available
+        run: |
+          which makeblastdb
+          which psiblast
+          which tblastn
+          makeblastdb -version
+          psiblast -version
+          tblastn -version
+          conda list -n sp4 | grep -E '^(blast|blast-legacy|selenoprofiles4)\s'
+
       - name: Run selenoprofiles test
         run: |
           echo "Running selenoprofiles test..."
           selenoprofiles -setup
           yes "" | selenoprofiles -download || true
-          selenoprofiles test
+          selenoprofiles test || {
+            status=$?
+            echo "::group::selenoprofiles test logs"
+            for log in log_test*; do
+              if [ -f "$log" ]; then
+                echo "===== $log ====="
+                tail -n 300 "$log"
+              fi
+            done
+            echo "::endgroup::"
+            exit "$status"
+          }
+
+      - name: Upload selenoprofiles test logs
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: selenoprofiles-test-logs
+          path: log_test*
+          if-no-files-found: ignore

Dockerfile

@@ -16,11 +16,6 @@ RUN apt-get update && apt-get install -y \
 
 RUN pip install selenoprofiles4[addons]
 
-ENV PATH=$PATH:/blast-2.2.26/bin
-
-RUN wget https://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/blast-2.2.26-x64-linux.tar.gz && \
-    tar -xzvf blast-2.2.26-x64-linux.tar.gz    
-
 # Run selenoprofiles setup and download commands
 RUN selenoprofiles -setup && \
     yes "" | selenoprofiles -download 
@@ -38,4 +33,7 @@ RUN conda install -c anaconda gawk
 
 RUN conda install bioconda::wise2
 
+# BLAST+ backend: provides makeblastdb, psiblast and tblastn.
+RUN conda install bioconda::blast
+
 RUN conda install -c mmariotti -c conda-forge -c etetoolkit ncbi_db

docs/installation.rst

@@ -44,6 +44,12 @@ Then **install selenoprofiles4** and its dependencies in the sp4 environment::
 
     conda install -c mmariotti -c anaconda  -c bioconda -c biobuilds selenoprofiles4
 
+Selenoprofiles 4.6 uses the BLAST+ package by default. If you install from
+source or maintain your own environment, make sure BLAST+ is present; it
+provides ``makeblastdb``, ``psiblast`` and ``tblastn``::
+
+    conda install -c bioconda blast
+
 If everything worked correctly, the selenoprofiles command is now available, but it is still not setup.
 Run this and follow instructions::
 
@@ -97,5 +103,9 @@ Check the :doc:`get_started` page to start using selenoprofiles.
 
         CONDA_SUBDIR=osx-64 conda install -c mmariotti -c anaconda -c bioconda -c biobuilds selenoprofiles4
 
-  After this, installation should succeed and *selenoprofiles4* will work normally on Apple Silicon.
+     For source/development installs, also ensure BLAST+ is installed in the
+     osx-64 environment::
 
+        CONDA_SUBDIR=osx-64 conda install -c bioconda blast
+
+  After this, installation should succeed and *selenoprofiles4* will work normally on Apple Silicon.

meta.yaml

@@ -23,7 +23,8 @@ requirements:
   run:
     - python >=3.12, <3.13
     - requests 
-    - bioconda::blast-legacy
+    # BLAST+ binaries: makeblastdb, psiblast and tblastn.
+    - bioconda::blast
     - bioconda::wise2
     - bioconda::exonerate
     - bioconda::mafft

src/selenoprofiles4/_version.py

@@ -1 +1 @@
-__version__="4.5.7"
+__version__="4.6.0"

src/selenoprofiles4/blaster_parser.awk

@@ -168,8 +168,11 @@ if (in_target_line){
 
 
 if ( /^>/)   {
-	if (FULL_TARGET){  subject = substr($0, 2)  }
-	else{ 	           subject = substr($1, 2) }
+	# BLAST+ may print subject lines as "> subject" (space after >),
+	# for which $1 is just ">". Parse from the full line first.
+	subject = substr($0, 2)
+	gsub(/^ +/, "", subject)
+	if (!FULL_TARGET){ split(subject, SUBJECT_SPLIT, " "); subject = SUBJECT_SPLIT[1] }
 
 	if (line_start==-1)
 	{line_start = NR }

src/selenoprofiles4/load_config.py

@@ -5,6 +5,7 @@
 selenoprofiles_data_dir = ~/selenoprofiles_data/
 ncpus = 1
 temp = /tmp/
+blast_backend = blastplus
 
 profiles_folder = {selenoprofiles_data_dir}/selenoprotein_profiles
 bin_folder =