Spaces:
No application file
No application file
Update app.py
Browse files
app.py
CHANGED
|
@@ -8,7 +8,7 @@ import os
|
|
| 8 |
import re
|
| 9 |
import logging
|
| 10 |
import numpy as np
|
| 11 |
-
from predictor import GenePredictor
|
| 12 |
from tensorflow.keras.models import load_model
|
| 13 |
import ml_simplified_tree
|
| 14 |
import tempfile
|
|
@@ -20,12 +20,13 @@ from tensorflow.keras.preprocessing.sequence import pad_sequences
|
|
| 20 |
|
| 21 |
# --- Global Variables ---
|
| 22 |
MAFFT_PATH = "mafft/mafftdir/bin/mafft" # Update this path as needed
|
| 23 |
-
IQTREE_PATH = "iqtree/bin/
|
| 24 |
|
| 25 |
# --- Logging ---
|
| 26 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
| 27 |
|
| 28 |
# --- Paths ---
|
|
|
|
| 29 |
model_repo = "GGproject10/best_boundary_aware_model"
|
| 30 |
csv_path = "f cleaned.csv"
|
| 31 |
classifier_model_dir = "model" # Directory for second model files
|
|
@@ -40,7 +41,6 @@ kmer_to_index = None
|
|
| 40 |
classifier_model = None
|
| 41 |
classifier_kmer_to_index = None
|
| 42 |
classifier_maxlen = None
|
| 43 |
-
LABELS = ["Random", "F", "P", "N", "M", "HN", "L"]
|
| 44 |
|
| 45 |
# Try to load boundary model from Hugging Face Hub
|
| 46 |
try:
|
|
@@ -56,7 +56,7 @@ except Exception as e:
|
|
| 56 |
# Try to load Keras model from Hugging Face Hub
|
| 57 |
try:
|
| 58 |
keras_path = hf_hub_download(repo_id=model_repo, filename="best_model.keras", token=hf_token)
|
| 59 |
-
kmer_path = hf_hub_download(repo_id=model_repo, filename="kmer_to_index.pkl")
|
| 60 |
if os.path.exists(keras_path) and os.path.exists(kmer_path):
|
| 61 |
keras_model = load_model(keras_path)
|
| 62 |
with open(kmer_path, "rb") as f:
|
|
@@ -73,25 +73,19 @@ try:
|
|
| 73 |
classifier_path = os.path.join(classifier_model_dir, "best_model.keras")
|
| 74 |
classifier_kmer_path = os.path.join(classifier_model_dir, "kmer_to_index.pkl")
|
| 75 |
classifier_maxlen_path = os.path.join(classifier_model_dir, "maxlen.txt")
|
| 76 |
-
|
| 77 |
-
if not os.path.exists(classifier_path):
|
| 78 |
-
missing_files.append("best_model.keras")
|
| 79 |
-
if not os.path.exists(classifier_kmer_path):
|
| 80 |
-
missing_files.append("kmer_to_index.pkl")
|
| 81 |
-
if not os.path.exists(classifier_maxlen_path):
|
| 82 |
-
missing_files.append("maxlen.txt")
|
| 83 |
-
if missing_files:
|
| 84 |
-
logging.warning(f"Classifier model files not found: {', '.join(missing_files)}")
|
| 85 |
-
else:
|
| 86 |
classifier_model = load_model(classifier_path)
|
| 87 |
with open(classifier_kmer_path, "rb") as f:
|
| 88 |
classifier_kmer_to_index = pickle.load(f)
|
| 89 |
with open(classifier_maxlen_path, "r") as f:
|
| 90 |
classifier_maxlen = int(f.read().strip())
|
| 91 |
logging.info("Classifier model loaded successfully.")
|
|
|
|
|
|
|
| 92 |
except Exception as e:
|
| 93 |
logging.error(f"Failed to load classifier model: {e}")
|
| 94 |
-
|
|
|
|
| 95 |
|
| 96 |
# --- Initialize Tree Analyzer ---
|
| 97 |
analyzer = None
|
|
@@ -133,12 +127,7 @@ def check_and_fix_executable_permissions(filepath):
|
|
| 133 |
return False
|
| 134 |
|
| 135 |
def enhanced_check_tool_availability():
|
| 136 |
-
"""Enhanced check for MAFFT and IQ-TREE availability with permission fixing
|
| 137 |
-
# Unset MAFFT_BINARIES to fix version check issue
|
| 138 |
-
if 'MAFFT_BINARIES' in os.environ:
|
| 139 |
-
del os.environ['MAFFT_BINARIES']
|
| 140 |
-
logging.info("Unset MAFFT_BINARIES environment variable to resolve version check issue.")
|
| 141 |
-
|
| 142 |
mafft_available = False
|
| 143 |
mafft_cmd = None
|
| 144 |
mafft_candidates = [
|
|
@@ -169,14 +158,15 @@ def enhanced_check_tool_availability():
|
|
| 169 |
iqtree_cmd = None
|
| 170 |
iqtree_candidates = [
|
| 171 |
IQTREE_PATH,
|
| 172 |
-
'
|
| 173 |
'iqtree',
|
| 174 |
-
'/usr/bin/
|
| 175 |
-
'/usr/local/bin/
|
| 176 |
'/usr/bin/iqtree',
|
| 177 |
'/usr/local/bin/iqtree',
|
| 178 |
-
'/opt/homebrew/bin/
|
| 179 |
-
'
|
|
|
|
| 180 |
]
|
| 181 |
for candidate in iqtree_candidates:
|
| 182 |
if candidate and os.path.exists(candidate):
|
|
@@ -208,7 +198,7 @@ def get_installation_instructions():
|
|
| 208 |
📦 INSTALLATION INSTRUCTIONS (Ubuntu/Debian):
|
| 209 |
1. Update package list: sudo apt-get update
|
| 210 |
2. Install MAFFT and IQ-TREE: sudo apt-get install mafft iqtree
|
| 211 |
-
3. Verify installation: mafft --version,
|
| 212 |
Alternative using Conda: conda install -c bioconda mafft iqtree
|
| 213 |
"""
|
| 214 |
elif 'centos' in os_info or 'rhel' in os_info or 'fedora' in os_info:
|
|
@@ -216,7 +206,7 @@ Alternative using Conda: conda install -c bioconda mafft iqtree
|
|
| 216 |
📦 INSTALLATION INSTRUCTIONS (CentOS/RHEL/Fedora):
|
| 217 |
1. Install EPEL repository (CentOS/RHEL): sudo yum install epel-release
|
| 218 |
2. Install packages: sudo yum install mafft iqtree
|
| 219 |
-
3. Verify installation: mafft --version,
|
| 220 |
"""
|
| 221 |
except:
|
| 222 |
pass
|
|
@@ -225,7 +215,7 @@ Alternative using Conda: conda install -c bioconda mafft iqtree
|
|
| 225 |
📦 INSTALLATION INSTRUCTIONS (macOS):
|
| 226 |
Using Homebrew: 1. Install Homebrew: /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
|
| 227 |
2. Install MAFFT and IQ-TREE: brew install mafft iqtree
|
| 228 |
-
3. Verify installation: mafft --version,
|
| 229 |
Using Conda: conda install -c bioconda mafft iqtree
|
| 230 |
"""
|
| 231 |
elif system == "windows":
|
|
@@ -251,14 +241,14 @@ def run_mafft_alignment_improved(input_fasta, output_fasta, mafft_cmd):
|
|
| 251 |
if not check_and_fix_executable_permissions(mafft_cmd):
|
| 252 |
return False, f"Cannot make {mafft_cmd} executable"
|
| 253 |
try:
|
| 254 |
-
test_result = subprocess.run([mafft_cmd, '--version'], capture_output=True, text=True, timeout=10
|
| 255 |
if test_result.returncode != 0:
|
| 256 |
return False, f"MAFFT version check failed: {test_result.stderr}"
|
| 257 |
except Exception as e:
|
| 258 |
return False, f"MAFFT version check failed: {str(e)}"
|
| 259 |
cmd = [mafft_cmd, '--auto', '--quiet', '--thread', '2', input_fasta]
|
| 260 |
logging.info(f"Running MAFFT: {' '.join(cmd)}")
|
| 261 |
-
result = subprocess.run(cmd, capture_output=True, text=True, timeout=600, cwd=os.getcwd()
|
| 262 |
if result.returncode == 0:
|
| 263 |
with open(output_fasta, 'w') as f:
|
| 264 |
f.write(result.stdout)
|
|
@@ -460,6 +450,77 @@ def analyze_sequence_for_tree(sequence: str, matching_percentage: float) -> str:
|
|
| 460 |
logging.error(f"Full traceback: {traceback.format_exc()}")
|
| 461 |
return error_msg
|
| 462 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 463 |
def read_fasta_file(file_obj):
|
| 464 |
"""Read FASTA file content"""
|
| 465 |
try:
|
|
@@ -483,30 +544,22 @@ def run_pipeline_from_file(fasta_file_obj, similarity_score, build_ml_tree):
|
|
| 483 |
dna_input = read_fasta_file(fasta_file_obj)
|
| 484 |
if not dna_input:
|
| 485 |
return "Failed to read FASTA file", "", "", "", "", "", "", "", "", None, None, None, "No input sequence"
|
| 486 |
-
return
|
| 487 |
-
dna_input, keras_model, kmer_to_index, classifier_model,
|
| 488 |
-
classifier_kmer_to_index, classifier_maxlen, LABELS,
|
| 489 |
-
similarity_score, build_ml_tree
|
| 490 |
-
)
|
| 491 |
except Exception as e:
|
| 492 |
error_msg = f"Pipeline error: {str(e)}"
|
| 493 |
logging.error(error_msg)
|
| 494 |
return error_msg, "", "", "", "", "", "", "", "", None, None, None, error_msg
|
| 495 |
|
| 496 |
-
def
|
| 497 |
-
|
| 498 |
-
similarity_score=95.0, build_ml_tree=False):
|
| 499 |
-
"""Enhanced pipeline with improved F gene prediction"""
|
| 500 |
try:
|
| 501 |
-
# Input validation and preprocessing
|
| 502 |
dna_input = dna_input.upper().strip()
|
| 503 |
if not dna_input:
|
| 504 |
return "Empty input", "", "", "", "", "", "", "", "", None, None, None, "No input provided"
|
| 505 |
-
|
| 506 |
if not re.match('^[ACTGN]+$', dna_input):
|
| 507 |
dna_input = ''.join(c if c in 'ACTGN' else 'N' for c in dna_input)
|
| 508 |
logging.info("DNA sequence sanitized")
|
| 509 |
-
|
| 510 |
# Step 1: Boundary Prediction
|
| 511 |
processed_sequence = dna_input
|
| 512 |
boundary_output = ""
|
|
@@ -516,7 +569,7 @@ def enhanced_run_pipeline(dna_input, keras_model, kmer_to_index, classifier_mode
|
|
| 516 |
regions = boundary_model.extract_gene_regions(predictions, dna_input)
|
| 517 |
if regions:
|
| 518 |
processed_sequence = regions[0]["sequence"]
|
| 519 |
-
boundary_output =
|
| 520 |
logging.info(f"F gene extracted: {len(processed_sequence)} bp (confidence: {confidence:.3f})")
|
| 521 |
else:
|
| 522 |
boundary_output = f"No F gene regions found in input sequence"
|
|
@@ -527,25 +580,18 @@ def enhanced_run_pipeline(dna_input, keras_model, kmer_to_index, classifier_mode
|
|
| 527 |
else:
|
| 528 |
boundary_output = f"Boundary model not available. Using original input: {len(dna_input)} bp"
|
| 529 |
|
| 530 |
-
# Step 2:
|
| 531 |
-
|
| 532 |
-
if
|
| 533 |
-
|
| 534 |
-
keras_output
|
| 535 |
-
keras_output += f"Sequence length: {keras_result['sequence_length']} nt"
|
| 536 |
-
if keras_result['kmer_coverage'] < 0.8:
|
| 537 |
-
keras_output += "\n⚠️ Low k-mer coverage - may affect accuracy"
|
| 538 |
-
else:
|
| 539 |
-
keras_output = str(keras_result)
|
| 540 |
|
| 541 |
-
# Step 3:
|
| 542 |
-
classifier_result =
|
| 543 |
-
processed_sequence, classifier_model, classifier_kmer_to_index, classifier_maxlen, labels
|
| 544 |
-
)
|
| 545 |
classifier_status = classifier_result["status"]
|
| 546 |
classifier_message = classifier_result["message"]
|
| 547 |
-
classifier_label = classifier_result["predicted_label"]
|
| 548 |
-
classifier_confidence =
|
| 549 |
|
| 550 |
# Step 4: Maximum Likelihood Tree
|
| 551 |
aligned_file = None
|
|
@@ -625,7 +671,7 @@ def enhanced_run_pipeline(dna_input, keras_model, kmer_to_index, classifier_mode
|
|
| 625 |
import traceback
|
| 626 |
logging.error(f"Full traceback: {traceback.format_exc()}")
|
| 627 |
return (
|
| 628 |
-
error_msg, "", "
|
| 629 |
None, None, None, error_msg
|
| 630 |
)
|
| 631 |
|
|
@@ -641,11 +687,11 @@ def create_interface():
|
|
| 641 |
gr.Markdown("""
|
| 642 |
# 🧬 F Gene Analysis Pipeline
|
| 643 |
|
| 644 |
-
This tool provides comprehensive analysis of F genes:
|
| 645 |
-
- **
|
| 646 |
-
- **
|
| 647 |
-
- **
|
| 648 |
-
- **
|
| 649 |
|
| 650 |
**Instructions:**
|
| 651 |
1. Enter your sequence or upload a FASTA file
|
|
@@ -693,9 +739,9 @@ def create_interface():
|
|
| 693 |
with gr.Row():
|
| 694 |
with gr.Column():
|
| 695 |
boundary_output = gr.Textbox(label="🎯 F Gene Extraction", lines=5, interactive=False)
|
| 696 |
-
keras_output = gr.Textbox(label="🔍 Gene Validation", lines=
|
| 697 |
classifier_status = gr.Textbox(label="🧬 Classification Status", lines=1, interactive=False)
|
| 698 |
-
classifier_message = gr.Textbox(label="📝 Classification Message", lines=
|
| 699 |
classifier_label = gr.Textbox(label="🏷️ Predicted Label", lines=1, interactive=False)
|
| 700 |
classifier_confidence = gr.Textbox(label="📊 Confidence Score", lines=1, interactive=False)
|
| 701 |
with gr.Column():
|
|
@@ -714,27 +760,27 @@ def create_interface():
|
|
| 714 |
## About This Tool
|
| 715 |
|
| 716 |
### F Gene Analysis Pipeline
|
| 717 |
-
- **🎯
|
| 718 |
- **🔍 Gene Validation**: Validates with k-mer based machine learning.
|
| 719 |
- **🧬 Gene Classification**: Classifies sequences (F gene or other) with confidence scores.
|
| 720 |
- **🌳 Phylogenetic Analysis**: Builds ML and simplified trees.
|
| 721 |
|
| 722 |
### Input Requirements
|
| 723 |
-
- DNA Sequences: ATCG format, minimum 50 bp
|
| 724 |
- FASTA Files: Standard format.
|
| 725 |
- Similarity Threshold: 1-99%.
|
| 726 |
|
| 727 |
### Dependencies
|
| 728 |
**For ML Trees:**
|
| 729 |
```bash
|
| 730 |
-
# Ubuntu/Debian: sudo apt-get
|
| 731 |
# macOS: brew install mafft iqtree
|
| 732 |
# Conda: conda install -c bioconda mafft iqtree
|
| 733 |
```
|
| 734 |
|
| 735 |
### Troubleshooting
|
| 736 |
- *"No similar sequences"*: Lower similarity threshold.
|
| 737 |
-
- *"Sequence too short"*: Provide >50 bp
|
| 738 |
- *"MAFFT/IQ-TREE not found"*: Install dependencies.
|
| 739 |
- *"Model not available"*: Check model files.
|
| 740 |
""")
|
|
@@ -743,11 +789,7 @@ def create_interface():
|
|
| 743 |
if file_obj is not None:
|
| 744 |
return run_pipeline_from_file(file_obj, sim_score, build_tree)
|
| 745 |
else:
|
| 746 |
-
return
|
| 747 |
-
dna_seq, keras_model, kmer_to_index, classifier_model,
|
| 748 |
-
classifier_kmer_to_index, classifier_maxlen, LABELS,
|
| 749 |
-
sim_score, build_tree
|
| 750 |
-
)
|
| 751 |
|
| 752 |
def clear_inputs():
|
| 753 |
return "", None, 95.0, False, "Ready to analyze"
|
|
|
|
| 8 |
import re
|
| 9 |
import logging
|
| 10 |
import numpy as np
|
| 11 |
+
from predictor import GenePredictor
|
| 12 |
from tensorflow.keras.models import load_model
|
| 13 |
import ml_simplified_tree
|
| 14 |
import tempfile
|
|
|
|
| 20 |
|
| 21 |
# --- Global Variables ---
|
| 22 |
MAFFT_PATH = "mafft/mafftdir/bin/mafft" # Update this path as needed
|
| 23 |
+
IQTREE_PATH = "iqtree/bin/iqtree2" # Update this path as needed
|
| 24 |
|
| 25 |
# --- Logging ---
|
| 26 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
| 27 |
|
| 28 |
# --- Paths ---
|
| 29 |
+
# Model repository and file paths
|
| 30 |
model_repo = "GGproject10/best_boundary_aware_model"
|
| 31 |
csv_path = "f cleaned.csv"
|
| 32 |
classifier_model_dir = "model" # Directory for second model files
|
|
|
|
| 41 |
classifier_model = None
|
| 42 |
classifier_kmer_to_index = None
|
| 43 |
classifier_maxlen = None
|
|
|
|
| 44 |
|
| 45 |
# Try to load boundary model from Hugging Face Hub
|
| 46 |
try:
|
|
|
|
| 56 |
# Try to load Keras model from Hugging Face Hub
|
| 57 |
try:
|
| 58 |
keras_path = hf_hub_download(repo_id=model_repo, filename="best_model.keras", token=hf_token)
|
| 59 |
+
kmer_path = hf_hub_download(repo_id=model_repo, filename="kmer_to_index.pkl", token=hf_token)
|
| 60 |
if os.path.exists(keras_path) and os.path.exists(kmer_path):
|
| 61 |
keras_model = load_model(keras_path)
|
| 62 |
with open(kmer_path, "rb") as f:
|
|
|
|
| 73 |
classifier_path = os.path.join(classifier_model_dir, "best_model.keras")
|
| 74 |
classifier_kmer_path = os.path.join(classifier_model_dir, "kmer_to_index.pkl")
|
| 75 |
classifier_maxlen_path = os.path.join(classifier_model_dir, "maxlen.txt")
|
| 76 |
+
if os.path.exists(classifier_path) and os.path.exists(classifier_kmer_path) and os.path.exists(classifier_maxlen_path):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
classifier_model = load_model(classifier_path)
|
| 78 |
with open(classifier_kmer_path, "rb") as f:
|
| 79 |
classifier_kmer_to_index = pickle.load(f)
|
| 80 |
with open(classifier_maxlen_path, "r") as f:
|
| 81 |
classifier_maxlen = int(f.read().strip())
|
| 82 |
logging.info("Classifier model loaded successfully.")
|
| 83 |
+
else:
|
| 84 |
+
logging.warning(f"Classifier model files not found in {classifier_model_dir}")
|
| 85 |
except Exception as e:
|
| 86 |
logging.error(f"Failed to load classifier model: {e}")
|
| 87 |
+
|
| 88 |
+
LABELS = ["Random", "F", "P", "N", "M", "HN", "L"]
|
| 89 |
|
| 90 |
# --- Initialize Tree Analyzer ---
|
| 91 |
analyzer = None
|
|
|
|
| 127 |
return False
|
| 128 |
|
| 129 |
def enhanced_check_tool_availability():
|
| 130 |
+
"""Enhanced check for MAFFT and IQ-TREE availability with permission fixing"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 131 |
mafft_available = False
|
| 132 |
mafft_cmd = None
|
| 133 |
mafft_candidates = [
|
|
|
|
| 158 |
iqtree_cmd = None
|
| 159 |
iqtree_candidates = [
|
| 160 |
IQTREE_PATH,
|
| 161 |
+
'iqtree2',
|
| 162 |
'iqtree',
|
| 163 |
+
'/usr/bin/iqtree2',
|
| 164 |
+
'/usr/local/bin/iqtree2',
|
| 165 |
'/usr/bin/iqtree',
|
| 166 |
'/usr/local/bin/iqtree',
|
| 167 |
+
'/opt/homebrew/bin/iqtree2',
|
| 168 |
+
'iqtree2.exe',
|
| 169 |
+
'iqtree.exe',
|
| 170 |
]
|
| 171 |
for candidate in iqtree_candidates:
|
| 172 |
if candidate and os.path.exists(candidate):
|
|
|
|
| 198 |
📦 INSTALLATION INSTRUCTIONS (Ubuntu/Debian):
|
| 199 |
1. Update package list: sudo apt-get update
|
| 200 |
2. Install MAFFT and IQ-TREE: sudo apt-get install mafft iqtree
|
| 201 |
+
3. Verify installation: mafft --version, iqtree2 --version
|
| 202 |
Alternative using Conda: conda install -c bioconda mafft iqtree
|
| 203 |
"""
|
| 204 |
elif 'centos' in os_info or 'rhel' in os_info or 'fedora' in os_info:
|
|
|
|
| 206 |
📦 INSTALLATION INSTRUCTIONS (CentOS/RHEL/Fedora):
|
| 207 |
1. Install EPEL repository (CentOS/RHEL): sudo yum install epel-release
|
| 208 |
2. Install packages: sudo yum install mafft iqtree
|
| 209 |
+
3. Verify installation: mafft --version, iqtree2 --version
|
| 210 |
"""
|
| 211 |
except:
|
| 212 |
pass
|
|
|
|
| 215 |
📦 INSTALLATION INSTRUCTIONS (macOS):
|
| 216 |
Using Homebrew: 1. Install Homebrew: /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
|
| 217 |
2. Install MAFFT and IQ-TREE: brew install mafft iqtree
|
| 218 |
+
3. Verify installation: mafft --version, iqtree2 --version
|
| 219 |
Using Conda: conda install -c bioconda mafft iqtree
|
| 220 |
"""
|
| 221 |
elif system == "windows":
|
|
|
|
| 241 |
if not check_and_fix_executable_permissions(mafft_cmd):
|
| 242 |
return False, f"Cannot make {mafft_cmd} executable"
|
| 243 |
try:
|
| 244 |
+
test_result = subprocess.run([mafft_cmd, '--version'], capture_output=True, text=True, timeout=10)
|
| 245 |
if test_result.returncode != 0:
|
| 246 |
return False, f"MAFFT version check failed: {test_result.stderr}"
|
| 247 |
except Exception as e:
|
| 248 |
return False, f"MAFFT version check failed: {str(e)}"
|
| 249 |
cmd = [mafft_cmd, '--auto', '--quiet', '--thread', '2', input_fasta]
|
| 250 |
logging.info(f"Running MAFFT: {' '.join(cmd)}")
|
| 251 |
+
result = subprocess.run(cmd, capture_output=True, text=True, timeout=600, cwd=os.getcwd())
|
| 252 |
if result.returncode == 0:
|
| 253 |
with open(output_fasta, 'w') as f:
|
| 254 |
f.write(result.stdout)
|
|
|
|
| 450 |
logging.error(f"Full traceback: {traceback.format_exc()}")
|
| 451 |
return error_msg
|
| 452 |
|
| 453 |
+
def predict_with_keras(sequence):
|
| 454 |
+
"""Keras prediction for initial sequence processing"""
|
| 455 |
+
try:
|
| 456 |
+
if not keras_model or not kmer_to_index:
|
| 457 |
+
return f"Keras model not available. Input sequence: {sequence[:100]}..."
|
| 458 |
+
if len(sequence) < 6:
|
| 459 |
+
return "Sequence too short for k-mer prediction (minimum 6 nucleotides required)."
|
| 460 |
+
kmers = [sequence[i:i+6] for i in range(len(sequence)-5)]
|
| 461 |
+
indices = [kmer_to_index.get(kmer, 0) for kmer in kmers]
|
| 462 |
+
input_arr = np.array([indices])
|
| 463 |
+
prediction = keras_model.predict(input_arr, verbose=0)[0]
|
| 464 |
+
result = ''.join([str(round(p, 3)) for p in prediction])
|
| 465 |
+
return result
|
| 466 |
+
except Exception as e:
|
| 467 |
+
logging.error(f"Keras prediction failed: {e}")
|
| 468 |
+
return f"Keras prediction failed: {str(e)}"
|
| 469 |
+
|
| 470 |
+
def classify_sequence(sequence):
|
| 471 |
+
"""Classify sequence using the second model"""
|
| 472 |
+
try:
|
| 473 |
+
if not classifier_model or not classifier_kmer_to_index or classifier_maxlen is None:
|
| 474 |
+
return {
|
| 475 |
+
"status": "error",
|
| 476 |
+
"message": "Classifier model not available.",
|
| 477 |
+
"confidence": None,
|
| 478 |
+
"predicted_label": None
|
| 479 |
+
}
|
| 480 |
+
if len(sequence) < 1500:
|
| 481 |
+
return {
|
| 482 |
+
"status": "error",
|
| 483 |
+
"message": "Sequence too short. Must be at least 1500 bases.",
|
| 484 |
+
"confidence": None,
|
| 485 |
+
"predicted_label": None
|
| 486 |
+
}
|
| 487 |
+
tokens = [sequence[i:i+6] for i in range(len(sequence)-5+1)]
|
| 488 |
+
encoded = [classifier_kmer_to_index.get(kmer, 0) for kmer in tokens]
|
| 489 |
+
padded = pad_sequences([encoded], maxlen=classifier_maxlen, padding='post')
|
| 490 |
+
pred = classifier_model.predict(padded, verbose=0)
|
| 491 |
+
predicted_class = int(np.argmax(pred))
|
| 492 |
+
label = LABELS[predicted_class]
|
| 493 |
+
confidence = float(np.max(pred))
|
| 494 |
+
if label == "F":
|
| 495 |
+
return {
|
| 496 |
+
"status": "success",
|
| 497 |
+
"message": "F gene detected.",
|
| 498 |
+
"confidence": confidence,
|
| 499 |
+
"predicted_label": label
|
| 500 |
+
}
|
| 501 |
+
elif label == "Random":
|
| 502 |
+
return {
|
| 503 |
+
"status": "error",
|
| 504 |
+
"message": "Unidentified sequence detected. Make sure you're entering the F gene of the NDV.",
|
| 505 |
+
"confidence": confidence,
|
| 506 |
+
"predicted_label": label
|
| 507 |
+
}
|
| 508 |
+
else:
|
| 509 |
+
return {
|
| 510 |
+
"status": "error",
|
| 511 |
+
"message": "No F-gene detected. Please enter an NDV's F gene.",
|
| 512 |
+
"confidence": confidence,
|
| 513 |
+
"predicted_label": label
|
| 514 |
+
}
|
| 515 |
+
except Exception as e:
|
| 516 |
+
logging.error(f"Classifier prediction failed: {e}")
|
| 517 |
+
return {
|
| 518 |
+
"status": "error",
|
| 519 |
+
"message": f"Prediction failed: {str(e)}",
|
| 520 |
+
"confidence": None,
|
| 521 |
+
"predicted_label": None
|
| 522 |
+
}
|
| 523 |
+
|
| 524 |
def read_fasta_file(file_obj):
|
| 525 |
"""Read FASTA file content"""
|
| 526 |
try:
|
|
|
|
| 544 |
dna_input = read_fasta_file(fasta_file_obj)
|
| 545 |
if not dna_input:
|
| 546 |
return "Failed to read FASTA file", "", "", "", "", "", "", "", "", None, None, None, "No input sequence"
|
| 547 |
+
return run_pipeline(dna_input, similarity_score, build_ml_tree)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 548 |
except Exception as e:
|
| 549 |
error_msg = f"Pipeline error: {str(e)}"
|
| 550 |
logging.error(error_msg)
|
| 551 |
return error_msg, "", "", "", "", "", "", "", "", None, None, None, error_msg
|
| 552 |
|
| 553 |
+
def run_pipeline(dna_input, similarity_score=95.0, build_ml_tree=False):
|
| 554 |
+
"""Run the full pipeline"""
|
|
|
|
|
|
|
| 555 |
try:
|
|
|
|
| 556 |
dna_input = dna_input.upper().strip()
|
| 557 |
if not dna_input:
|
| 558 |
return "Empty input", "", "", "", "", "", "", "", "", None, None, None, "No input provided"
|
|
|
|
| 559 |
if not re.match('^[ACTGN]+$', dna_input):
|
| 560 |
dna_input = ''.join(c if c in 'ACTGN' else 'N' for c in dna_input)
|
| 561 |
logging.info("DNA sequence sanitized")
|
| 562 |
+
|
| 563 |
# Step 1: Boundary Prediction
|
| 564 |
processed_sequence = dna_input
|
| 565 |
boundary_output = ""
|
|
|
|
| 569 |
regions = boundary_model.extract_gene_regions(predictions, dna_input)
|
| 570 |
if regions:
|
| 571 |
processed_sequence = regions[0]["sequence"]
|
| 572 |
+
boundary_output = processed_sequence
|
| 573 |
logging.info(f"F gene extracted: {len(processed_sequence)} bp (confidence: {confidence:.3f})")
|
| 574 |
else:
|
| 575 |
boundary_output = f"No F gene regions found in input sequence"
|
|
|
|
| 580 |
else:
|
| 581 |
boundary_output = f"Boundary model not available. Using original input: {len(dna_input)} bp"
|
| 582 |
|
| 583 |
+
# Step 2: Keras Prediction
|
| 584 |
+
keras_output = ""
|
| 585 |
+
if processed_sequence and len(processed_sequence) >= 6:
|
| 586 |
+
keras_prediction = predict_with_keras(processed_sequence)
|
| 587 |
+
keras_output = keras_prediction if not keras_prediction.startswith(("Keras", "Sequence")) else keras_prediction
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 588 |
|
| 589 |
+
# Step 3: Classifier Prediction
|
| 590 |
+
classifier_result = classify_sequence(processed_sequence)
|
|
|
|
|
|
|
| 591 |
classifier_status = classifier_result["status"]
|
| 592 |
classifier_message = classifier_result["message"]
|
| 593 |
+
classifier_label = classifier_result["predicted_label"]
|
| 594 |
+
classifier_confidence = classifier_result["confidence"]
|
| 595 |
|
| 596 |
# Step 4: Maximum Likelihood Tree
|
| 597 |
aligned_file = None
|
|
|
|
| 671 |
import traceback
|
| 672 |
logging.error(f"Full traceback: {traceback.format_exc()}")
|
| 673 |
return (
|
| 674 |
+
error_msg, "", "", "", "", "", "", "", f"<p>Error: {error_msg}</p>",
|
| 675 |
None, None, None, error_msg
|
| 676 |
)
|
| 677 |
|
|
|
|
| 687 |
gr.Markdown("""
|
| 688 |
# 🧬 F Gene Analysis Pipeline
|
| 689 |
|
| 690 |
+
This tool provides comprehensive analysis of F genes including:
|
| 691 |
+
- **Gene Boundary Detection**: Extract F gene sequences
|
| 692 |
+
- **Gene Validation**: Validate with machine learning
|
| 693 |
+
- **Gene Classification**: Classify sequence type (F gene or other)
|
| 694 |
+
- **Phylogenetic Analysis**: Build maximum likelihood and simplified trees
|
| 695 |
|
| 696 |
**Instructions:**
|
| 697 |
1. Enter your sequence or upload a FASTA file
|
|
|
|
| 739 |
with gr.Row():
|
| 740 |
with gr.Column():
|
| 741 |
boundary_output = gr.Textbox(label="🎯 F Gene Extraction", lines=5, interactive=False)
|
| 742 |
+
keras_output = gr.Textbox(label="🔍 Gene Validation", lines=3, interactive=False)
|
| 743 |
classifier_status = gr.Textbox(label="🧬 Classification Status", lines=1, interactive=False)
|
| 744 |
+
classifier_message = gr.Textbox(label="📝 Classification Message", lines=2, interactive=False)
|
| 745 |
classifier_label = gr.Textbox(label="🏷️ Predicted Label", lines=1, interactive=False)
|
| 746 |
classifier_confidence = gr.Textbox(label="📊 Confidence Score", lines=1, interactive=False)
|
| 747 |
with gr.Column():
|
|
|
|
| 760 |
## About This Tool
|
| 761 |
|
| 762 |
### F Gene Analysis Pipeline
|
| 763 |
+
- **🎯 Gene Boundary Detection**: Extracts F gene sequences using deep learning.
|
| 764 |
- **🔍 Gene Validation**: Validates with k-mer based machine learning.
|
| 765 |
- **🧬 Gene Classification**: Classifies sequences (F gene or other) with confidence scores.
|
| 766 |
- **🌳 Phylogenetic Analysis**: Builds ML and simplified trees.
|
| 767 |
|
| 768 |
### Input Requirements
|
| 769 |
+
- DNA Sequences: ATCG format, minimum 50 bp.
|
| 770 |
- FASTA Files: Standard format.
|
| 771 |
- Similarity Threshold: 1-99%.
|
| 772 |
|
| 773 |
### Dependencies
|
| 774 |
**For ML Trees:**
|
| 775 |
```bash
|
| 776 |
+
# Ubuntu/Debian: sudo apt-get install mafft iqtree
|
| 777 |
# macOS: brew install mafft iqtree
|
| 778 |
# Conda: conda install -c bioconda mafft iqtree
|
| 779 |
```
|
| 780 |
|
| 781 |
### Troubleshooting
|
| 782 |
- *"No similar sequences"*: Lower similarity threshold.
|
| 783 |
+
- *"Sequence too short"*: Provide >50 bp.
|
| 784 |
- *"MAFFT/IQ-TREE not found"*: Install dependencies.
|
| 785 |
- *"Model not available"*: Check model files.
|
| 786 |
""")
|
|
|
|
| 789 |
if file_obj is not None:
|
| 790 |
return run_pipeline_from_file(file_obj, sim_score, build_tree)
|
| 791 |
else:
|
| 792 |
+
return run_pipeline(dna_seq, sim_score, build_tree)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 793 |
|
| 794 |
def clear_inputs():
|
| 795 |
return "", None, 95.0, False, "Ready to analyze"
|