YuuMJ1997/EukPhylo
1
0
Fork 0
mirror of http://43.156.76.180:8026/YuuMJ/EukPhylo.git synced 2025-12-27 04:00:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update CladeGrabbing.py

Browse Source 
Adding new optional arguments to support two-step clade filtering within one run of the script.
This commit is contained in:
Adri K. Grow 2025-08-18 10:54:25 -04:00 committed by GitHub
parent 873862adbc
commit 20d3926f10
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 137 additions and 66 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

95
Utilities/for_trees/CladeGrabbing.py
View File

@ -1,10 +1,10 @@
#Author, date: Auden Cote-L'Heureux, last updated Apr 1st 2024 by GA
#Author, date: Auden Cote-L'Heureux, last updated Aug 18th 2025 by AKG
#Motivation: Select robust sequences from trees
#Intent: Select clades of interest from large trees using taxonomic specifications
#Dependencies: Python3, ete3, Biopython
#Inputs: A folder containing: all PTLp2 output trees and all corresponding unaligned .fasta (pre-guidance) files
#Outputs: A folder of grabbed clades and filtered unaligned fasta files
#Example: python CladeGrabbing.py --input /Path/To/TreesandPreGuidance --target Sr_rh --min_presence 20
#Example: python3 CladeGrabbing.py --input /Path/To/TreesandPreGuidance --target Sr_rh --min_presence 20
#IMPORTANT: key parameters explained in "add_argument" section below
#Dependencies
@ -28,6 +28,8 @@ def get_args():
parser.add_argument('-nr', '--required_taxa_num', type = int, default = 0, help = 'The number of species belonging to taxa in the --required_taxa list that must be present in the clade. Default is 0.')
parser.add_argument('-o', '--outgroup', type = str, default = '', help = 'A comma-separated list of any number of digits/characters (e.g. Sr_ci_S OR Am_t), or a file with the extension .txt containing a list of complete or partial taxon codes, to describe taxa that will be included as outgroups in the output unaligned fasta files (which will contain only sequences from a single selected clade, and all outgroup sequences in the tree captured by this argument).')
parser.add_argument('-c', '--contaminants', type = float, default = 2, help = 'The number of non-ingroup contaminants allowed in a clade, or if less than 1 the proportion of sequences in a clade that can be non-ingroup (i.e. presumed contaminants). Default is to allow 2 contaminants.')
parser.add_argument('-ft', '--first_target', type=str, default='', help='[Optional] A comma-separated list or .txt file of complete/partial taxon codes for an initial, broad clade search. If provided, the script will first find clades with these taxa before applying the main --target filter.')
parser.add_argument('-fm', '--first_min_presence', type=int, default=0, help='[Optional] Minimum number of sequences from --first_target required in a clade for it to be used in the second-stage search. Ignored if --first_target is not provided.')
return parser.parse_args()
@ -86,15 +88,82 @@ def reroot(tree):
def get_subtrees(args, file):
newick = get_newick(args.input + '/' + file)
tree = ete3.Tree(newick)
majs = list(dict.fromkeys([leaf.name[:2] for leaf in tree]))
#Only try to reroot trees with more than 2 major clades. This was added to fix the ETE3 "Cannot set myself as outgroup" error
# Only try to reroot trees with more than 2 major clades (original behavior)
if len(majs) > 2:
tree = reroot(tree)
# -------------------------------
# FIRST-STAGE (optional) FILTER
# -------------------------------
def get_outer_leafsets():
"""
Return a list of sets, each set = leaf names of an outer clade
that passes --first_target, --first_min_presence, children_keep,
and contaminants logic (using args.contaminants).
If --first_target is not used, return one set containing ALL leaves.
"""
if not args.first_target or args.first_min_presence == 0:
return [set(leaf.name for leaf in tree)] # no outer filter → whole tree
# Parse first_target codes
if '.' in args.first_target:
first_target_codes = [l.strip() for l in open(args.first_target, 'r').readlines() if l.strip() != '']
else:
first_target_codes = [code.strip() for code in args.first_target.split(',') if code.strip() != '']
outer_sets = []
seen_leaves = []
for node in tree.traverse('levelorder'):
# large enough and not subsumed by already accepted outer node
if len(node) >= args.first_min_presence and len(set(seen_leaves) & set([leaf.name for leaf in node])) == 0:
leaves = [leaf.name for leaf in node]
# children_keep logic but for first_target
children_keep = 0
for child in node.children:
taken = False
for code in first_target_codes:
for leaf in child:
if leaf.name.startswith(code):
children_keep += 1
taken = True
break
if taken:
break
if children_keep != len(node.children):
continue
# count first-target hits (use [:10] uniqueness like original)
first_hits = set()
for code in first_target_codes:
for leaf in leaves[::-1]:
if leaf.startswith(code):
first_hits.add(leaf[:10])
leaves.remove(leaf)
# contaminants logic applied to FIRST-STAGE (reuse args.contaminants)
passes_contam = ((args.contaminants < 1 and len(leaves) <= args.contaminants * len(first_hits)) or
(args.contaminants >= 1 and len(leaves) <= args.contaminants))
if len(first_hits) >= args.first_min_presence and passes_contam:
outer_sets.append(set(leaf.name for leaf in node))
seen_leaves.extend([leaf.name for leaf in node])
return outer_sets
# Build outer sets; if user supplied first-stage args, we'll restrict inner search to these
using_first = bool(args.first_target) and args.first_min_presence > 0
outer_leafsets = get_outer_leafsets()
# --------------------------------
# ORIGINAL INNER FILTER (unchanged)
# --------------------------------
# Getting a clean list of all target taxa
if '.' in args.target:
try:
@ -115,18 +184,23 @@ def get_subtrees(args, file):
target_codes = list(dict.fromkeys(target_codes + required_taxa_codes))
# Creating a record of selected subtrees, and all of the leaves in those subtrees
selected_nodes = []; seen_leaves = []
# Iterating through all nodes in tree, starting at "root" then working towards leaves
for node in tree.traverse('levelorder'):
#If a node is large enough and is not contained in an already selected clade
# If using first-stage filter, only consider nodes fully inside some outer clade
if using_first:
node_leafs = set(leaf.name for leaf in node)
# require subset (node fully contained in an accepted outer clade)
if not any(node_leafs.issubset(S) for S in outer_leafsets):
continue
# If a node is large enough and is not contained in an already selected clade
if len(node) >= args.min_presence and len(list(set(seen_leaves) & set([leaf.name for leaf in node]))) == 0:
leaves = [leaf.name for leaf in node]
#Accounting for cases where e.g. one child is a contaminant, and the other child is a good clade with 1 fewer than the max number of contaminants
# Accounting for cases where e.g. one child is a contaminant, and the other child is a good clade
children_keep = 0
for child in node.children:
for code in target_codes:
@ -144,7 +218,6 @@ def get_subtrees(args, file):
for code in target_codes:
for leaf in leaves[::-1]:
#print(leaf)
if leaf.startswith(code):
target_leaves.add(leaf[:10])
@ -154,13 +227,11 @@ def get_subtrees(args, file):
break
leaves.remove(leaf)
#Grab a clade as a subtree if 1) it has enough target taxa; 2) it has enough "at least" taxa; 3) it does not have too many contaminants
# Grab a clade as a subtree if it passes all filters
if len(target_leaves) >= args.min_presence and len(required_taxa_leaves) >= args.required_taxa_num and ((args.contaminants < 1 and len(leaves) <= args.contaminants * len(target_leaves)) or len(leaves) <= args.contaminants):
selected_nodes.append(node)
seen_leaves.extend([leaf.name for leaf in node])
# Write the subtrees to output .tre files
for i, node in enumerate(selected_nodes[::-1]):
with open('Subtrees/' + '.'.join(file.split('.')[:-1]) + '_' + str(i) + '.tre', 'w') as o: