Use pyHMMER to get rid of binary dependency on HMMER

.github/workflows/build-and-test.yml

@@ -0,0 +1,27 @@
+name: Build & Test
+
+on:
+  push:
+  schedule:
+  - cron:  '0 3 * * 1'
+
+
+jobs:
+  build-linux:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python
+      uses: actions/setup-python@v2
+      with:
+        python-version: 3.8
+    - name: Install dependencies
+      run: |
+        pip install -r requirements.txt
+    - name: Install ANARCI
+      run: |
+        python setup.py install
+    - name: Test with pytest
+      run: |
+        pip install pytest
+        pytest ./test

bin/ANARCI

@@ -147,16 +147,8 @@ if __name__ == "__main__":
 
     # Check that hmmscan can be found in the path
     if args.hmmerpath:
-        hmmerpath=args.hmmerpath
-        scan_path = os.path.join( hmmerpath, "hmmscan" )
-        if not ( os.path.exists( scan_path ) and os.access(scan_path, os.X_OK)):
-            print("Error: No hmmscan executable file found in directory: %s"%(hmmerpath), file=sys.stderr)
-            sys.exit(1)
-    elif not which("hmmscan"):
-        print("Error: hmmscan was not found in the path. Either install and add to path or provide path with commandline option.", file=sys.stderr)
-        sys.exit(1)
-
-
+        print('Note: ANARCI is using pyhmmer, ignoring --hmmerpath argument')
+
     # Check if there should be some restriction as to which chain types should be numbered.
     # If it is not the imgt scheme they want then restrict to only igs (otherwise you'll hit assertion errors)
     types_to_chains = {"ig":["H","K","L"], "tr":["A", "B","G","D"], "heavy":["H"], "light":["K","L"] }

build_pipeline/BuildHMM.py

@@ -0,0 +1,31 @@
+"""
+Build All.HMM from stockholm aligned file using PyHMMER
+"""
+
+import sys
+import pyhmmer
+
+
+def build(out_path, in_path):
+    alphabet = pyhmmer.easel.Alphabet.amino()
+    print('Loading MSA from {}'.format(in_path))
+    with pyhmmer.easel.MSAFile(in_path, digital=True, alphabet=alphabet) as msa_file:
+        msas = list(msa_file)
+
+    builder = pyhmmer.plan7.Builder(alphabet, architecture='hand')
+    background = pyhmmer.plan7.Background(alphabet)
+
+    print('Building HMMs from {} MSAs'.format(len(msas)))
+    hmms = []
+    with open(out_path, "wb") as output_file:
+        for msa in msas:
+            hmm, _, _ = builder.build_msa(msa, background)
+            hmm.write(output_file)
+            hmms.append(hmm)
+    print('Pressing HMMs')
+    pyhmmer.hmmpress(hmms, out_path)
+    print('Saved HMMs to {}'.format(out_path))
+
+
+if __name__ == '__main__':
+    build(sys.argv[1], sys.argv[2])

build_pipeline/RUN_pipeline.sh

@@ -15,10 +15,4 @@ python3 $DIR/FormatAlignments.py
 # Build the hmms for each species and chain.
 # --hand option required otherwise it will delete columns that are mainly gaps. We want 128 columns otherwise ARNACI will fall over.
 mkdir -p $DIR/HMMs
-hmmbuild --hand $DIR/HMMs/ALL.hmm $DIR/curated_alignments/ALL.stockholm
-#hmmbuild --hand $DIR/HMMs/ALL_AND_C.hmm $DIR/curated_alignments/ALL_AND_C.stockholm
-
-# Turn the output HMMs file into a binary form. This is required for hmmscan that is used in ARNACI.
-hmmpress -f $DIR/HMMs/ALL.hmm 
-#hmmpress -f $DIR/HMMs/ALL_AND_C.hmm
-
+python3 $DIR/BuildHMM.py $DIR/HMMs/ALL.hmm $DIR/curated_alignments/ALL.stockholm