Aligner extra args

CHANGES.md

@@ -2,6 +2,7 @@
 
 ## __NEXT__
 
+* align: Added `--alignment-args` options for passing arguments to the alignment program. [#1789] (@vbadelita)
 
 ## 30.0.0 (15 April 2025)
 

augur/align.py

@@ -14,6 +14,10 @@
 from .utils import nthreads_value
 from collections import defaultdict
 
+DEFAULT_ARGS = {
+    "mafft": "--reorder --anysymbol --nomemsave --adjustdirection",
+}
+
 class AlignmentError(Exception):
     # TODO: this exception should potentially be renamed and made augur-wide
     # thus allowing any module to raise it and have the message printed & augur
@@ -31,6 +35,8 @@ def register_arguments(parser):
     parser.add_argument('--nthreads', type=nthreads_value, default=1,
                                 help="number of threads to use; specifying the value 'auto' will cause the number of available CPU cores on your system, if determinable, to be used")
     parser.add_argument('--method', default='mafft', choices=["mafft"], help="alignment program to use")
+    parser.add_argument('--alignment-args', help="arguments to pass to the alignment program (except for threads, keeplength if --existing-alignment is passed), overriding defaults. " +
+                                                f"mafft defaults: '{DEFAULT_ARGS['mafft']}'")
     parser.add_argument('--reference-name', metavar="NAME", type=str, help="strip insertions relative to reference sequence; use if the reference is already in the input sequences")
     parser.add_argument('--reference-sequence', metavar="PATH", type=str, help="Add this reference sequence to the dataset & strip insertions relative to this. Use if the reference is NOT already in the input sequences")
     parser.add_argument('--remove-reference', action="store_true", default=False, help="remove reference sequence from the alignment")
@@ -132,7 +138,7 @@ def run(args):
 
         # generate alignment command & run
         log = args.output + ".log"
-        cmd = generate_alignment_cmd(args.method, args.nthreads, existing_aln_fname, seqs_to_align_fname, args.output, log)
+        cmd = generate_alignment_cmd(args.method, args.nthreads, existing_aln_fname, seqs_to_align_fname, args.output, log, alignment_args=args.alignment_args)
         success = run_shell_command(cmd)
         if not success:
             raise AlignmentError(f"Error during alignment: please see the log file {log!r} for more details")
@@ -248,17 +254,28 @@ def read_reference(ref_fname):
                 "\n\tmake sure the file %s contains one sequence in genbank or fasta format"%ref_fname)
     return ref_seq
 
-def generate_alignment_cmd(method, nthreads, existing_aln_fname, seqs_to_align_fname, aln_fname, log_fname):
+def generate_alignment_cmd(method, nthreads, existing_aln_fname, seqs_to_align_fname, aln_fname, log_fname, alignment_args):
+    if method not in DEFAULT_ARGS:
+        raise AlignmentError('ERROR: alignment method %s not implemented'%method)
+
+    if alignment_args is None:
+        alignment_args = DEFAULT_ARGS[method]
+
     if method=='mafft':
+        files_to_align = shquote(seqs_to_align_fname)
         if existing_aln_fname:
-            cmd = "mafft --add %s --keeplength --reorder --anysymbol --nomemsave --adjustdirection --thread %d %s 1> %s 2> %s"%(shquote(seqs_to_align_fname), nthreads, shquote(existing_aln_fname), shquote(aln_fname), shquote(log_fname))
-        else:
-            cmd = "mafft --reorder --anysymbol --nomemsave --adjustdirection --thread %d %s 1> %s 2> %s"%(nthreads, shquote(seqs_to_align_fname), shquote(aln_fname), shquote(log_fname))
+            # If there is an existing alignment, then seqs_to_align_fname becomes a parameter of --add
+            # and existing_aln_fname becomes the anonymous parameter
+            files_to_align = f"--add {shquote(seqs_to_align_fname)} {shquote(existing_aln_fname)}"
+            alignment_args = " ".join(["--keeplength", alignment_args])
+
+        cmd = f"mafft {alignment_args} --thread {nthreads} {files_to_align} 1> {shquote(aln_fname)} 2> {shquote(log_fname)}"
         print("\nusing mafft to align via:\n\t" + cmd +
               " \n\n\tKatoh et al, Nucleic Acid Research, vol 30, issue 14"
               "\n\thttps://doi.org/10.1093%2Fnar%2Fgkf436\n")
     else:
         raise AlignmentError('ERROR: alignment method %s not implemented'%method)
+
     return cmd
 
 

tests/test_align.py

@@ -187,7 +187,7 @@ def test_prettify_alignment(self):
 
     def test_generate_alignment_cmd_non_mafft(self):
         with pytest.raises(align.AlignmentError):
-            assert align.generate_alignment_cmd('no-mafft', 1, None, None, None, None)
+            assert align.generate_alignment_cmd('no-mafft', 1, None, None, None, None, alignment_args=None)
 
     def test_generate_alignment_cmd_mafft_existing_aln_fname(self):
         existing_aln_fname = "existing_aln"
@@ -199,9 +199,10 @@ def test_generate_alignment_cmd_mafft_existing_aln_fname(self):
                                               existing_aln_fname,
                                               seqs_to_align_fname,
                                               aln_fname,
-                                              log_fname)
+                                              log_fname,
+                                              alignment_args=None)
 
-        expected = "mafft --add %s --keeplength --reorder --anysymbol --nomemsave --adjustdirection --thread %d %s 1> %s 2> %s" % (quote(seqs_to_align_fname), 1, quote(existing_aln_fname), quote(aln_fname), quote(log_fname))
+        expected = "mafft --keeplength --reorder --anysymbol --nomemsave --adjustdirection --thread %d --add %s %s 1> %s 2> %s" % (1, quote(seqs_to_align_fname), quote(existing_aln_fname), quote(aln_fname), quote(log_fname))
 
         assert result == expected
 
@@ -214,11 +215,45 @@ def test_generate_alignment_cmd_mafft_no_existing_aln_fname(self):
                                               None,
                                               seqs_to_align_fname,
                                               aln_fname,
-                                              log_fname)
+                                              log_fname,
+                                              alignment_args=None)
 
         expected = "mafft --reorder --anysymbol --nomemsave --adjustdirection --thread %d %s 1> %s 2> %s" % (1, quote(seqs_to_align_fname), quote(aln_fname), quote(log_fname))
 
         assert result == expected
+
+    def test_generate_alignment_cmd_mafft_custom_args_existing_aln_fname(self):
+        existing_aln_fname = "existing_aln"
+        seqs_to_align_fname = "seqs_to_align"
+        aln_fname = "aln_fname"
+        log_fname = "log_fname"
+
+        result = align.generate_alignment_cmd("mafft", 1,
+                                              existing_aln_fname,
+                                              seqs_to_align_fname,
+                                              aln_fname,
+                                              log_fname,
+                                              alignment_args="--auto")
+
+        expected = "mafft --keeplength --auto --thread %d --add %s %s 1> %s 2> %s" % (1, quote(seqs_to_align_fname), quote(existing_aln_fname), quote(aln_fname), quote(log_fname))
+
+        assert result == expected
+
+    def test_generate_alignment_cmd_mafft_custom_args_no_existing_aln_fname(self):
+        seqs_to_align_fname = "seqs_to_align"
+        aln_fname = "aln_fname"
+        log_fname = "log_fname"
+
+        result = align.generate_alignment_cmd("mafft", 1,
+                                              None,
+                                              seqs_to_align_fname,
+                                              aln_fname,
+                                              log_fname,
+                                              alignment_args="--auto --anysymbol")
+
+        expected = "mafft --auto --anysymbol --thread %d %s 1> %s 2> %s" % (1, quote(seqs_to_align_fname), quote(aln_fname), quote(log_fname))
+
+        assert result == expected
 
     def test_read_alignment(self):
         data_file = pathlib.Path('tests/data/align/test_aligned_sequences.fasta')