phykit 2.0.0__tar.gz → 2.0.1__tar.gz

{phykit-2.0.0 → phykit-2.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: phykit
-Version: 2.0.0
+Version: 2.0.1
 Home-page: https://github.com/jlsteenwyk/phykit
 Author: Jacob L. Steenwyk
 Author-email: jlsteenwyk@gmail.com

{phykit-2.0.0 → phykit-2.0.1}/phykit/phykit.py

@@ -2475,6 +2475,8 @@ class Phykit(object):
                 
                 -t/--tree                   a tree file
 
+                -e/--exclude_gaps           if a site has a gap, ignore it
+
                 -v/--verbose                print out patristic distances
                                             and uncorrected distances used
                                             to determine saturation
@@ -2490,6 +2492,9 @@ class Phykit(object):
         parser.add_argument(
             "-v", "--verbose", action="store_true", required=False, help=SUPPRESS
         )
+        parser.add_argument(
+            "-e", "--exclude_gaps", action="store_true", required=False, help=SUPPRESS
+        )
         args = parser.parse_args(argv)
         Saturation(args).run()
 

{phykit-2.0.0 → phykit-2.0.1}/phykit/services/alignment/pairwise_identity.py

@@ -57,13 +57,18 @@ class PairwiseIdentity(Alignment):
             total_compared = 0
 
             for res_one, res_two in zip(seq_one, seq_two):
-                if exclude_gaps and (
-                    res_one not in gap_chars or res_two not in gap_chars
-                ):
-                    continue
+                res_one = res_one.upper()
+                res_two = res_two.upper()
+
                 total_compared += 1
-                if res_one == res_two:
-                    identities += 1
+
+                if exclude_gaps:
+                    if (res_one not in gap_chars or res_two not in gap_chars):
+                        if res_one == res_two:
+                            identities += 1
+                else:
+                    if res_one == res_two:
+                        identities += 1
 
             if total_compared > 0:
                 identity_score = identities / total_compared

{phykit-2.0.0 → phykit-2.0.1}/phykit/services/tree/base.py

@@ -1,4 +1,5 @@
 import sys
+from typing import List
 
 from Bio import Phylo
 
@@ -27,6 +28,7 @@ class Tree(BaseService):
         tip_2=None,
         clade=None,
         keep=None,
+        exclude_gaps=None,
     ):
         self.tree_file_path = tree_file_path
         self.tree1_file_path = tree1_file_path
@@ -47,6 +49,7 @@ class Tree(BaseService):
         self.tip_2 = tip_2
         self.clade = clade
         self.keep = keep
+        self.exclude_gaps = exclude_gaps
 
     def read_tree_file(self):
         try:
@@ -141,3 +144,9 @@ class Tree(BaseService):
                 return None
             except BrokenPipeError:
                 pass
+
+    def get_gap_chars(is_protein: bool) -> List[str]:
+        if is_protein:
+            return ["-", "?", "*", "X", "x"]
+        else:
+            return ["-", "?", "*", "X", "x", "N", "n"]

{phykit-2.0.0 → phykit-2.0.1}/phykit/services/tree/saturation.py

@@ -30,7 +30,7 @@ class Saturation(Tree):
         super().__init__(**self.process_args(args))
 
     def run(self) -> None:
-        alignment, _, _ = get_alignment_and_format_helper(
+        alignment, _, is_protein = get_alignment_and_format_helper(
             self.alignment_file_path
         )
 
@@ -43,7 +43,7 @@ class Saturation(Tree):
             patristic_distances,
             uncorrected_distances,
         ) = self.loop_through_combos_and_calculate_pds_and_pis(
-            combos, alignment, tree
+            combos, alignment, tree, self.exclude_gaps
         )
 
         # calculate slope and fit the y-intercept to zero
@@ -63,6 +63,7 @@ class Saturation(Tree):
         return dict(
             tree_file_path=args.tree,
             alignment_file_path=args.alignment,
+            exclude_gaps=args.exclude_gaps,
             verbose=args.verbose,
         )
 
@@ -71,6 +72,7 @@ class Saturation(Tree):
         combos: List[Tuple[str, str]],
         alignment: Align.MultipleSeqAlignment,
         tree: Newick.Tree,
+        exclude_gaps: bool,
     ) -> Tuple[
         List[float],
         List[float]
@@ -81,6 +83,7 @@ class Saturation(Tree):
         """
         patristic_distances = []
         uncorrected_distances = []
+        gap_chars = self.get_gap_chars()
         aln_len = alignment.get_alignment_length()
         seq_dict = {record.name: record.seq for record in alignment}
         for combo in combos:
@@ -90,8 +93,26 @@ class Saturation(Tree):
             # calcualte uncorrected distances
             seq_one = seq_dict[combo[0]]
             seq_two = seq_dict[combo[1]]
-            identities = sum(1 for idx in range(aln_len) if seq_one[idx] == seq_two[idx])
-            uncorrected_distances.append(1 - (identities / aln_len))
+            if exclude_gaps:
+                valid_positions = [
+                    idx for idx in range(aln_len)
+                    if seq_one[idx] not in gap_chars and seq_two[idx] not in gap_chars
+                ]
+                adjusted_len = len(valid_positions)
+                identities = sum(
+                    1 for idx in valid_positions if seq_one[idx].upper() == seq_two[idx].upper()
+                )
+            else:
+                adjusted_len = aln_len
+                identities = sum(
+                    1 for idx in range(aln_len) if seq_one[idx].upper() == seq_two[idx].upper()
+                )
+
+            if adjusted_len > 0:
+                uncorrected_distances.append(1 - (identities / adjusted_len))
+            else:
+                uncorrected_distances.append(float('nan'))
+                uncorrected_distances.append(1 - (identities / aln_len))
 
         return patristic_distances, uncorrected_distances
 

phykit-2.0.1/phykit/version.py

@@ -0,0 +1 @@
+__version__ = "2.0.1"

{phykit-2.0.0 → phykit-2.0.1}/phykit.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: phykit
-Version: 2.0.0
+Version: 2.0.1
 Home-page: https://github.com/jlsteenwyk/phykit
 Author: Jacob L. Steenwyk
 Author-email: jlsteenwyk@gmail.com

phykit-2.0.0/phykit/version.py

@@ -1 +0,0 @@
-__version__ = "2.0.0"