File size: 10,472 Bytes
4636192 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 | """Phase 2 v2: Parse raw sample CSVs and match against reference genotypes - IMPROVED MATCHING."""
import pandas as pd
import numpy as np
from pathlib import Path
import glob
import json
from collections import defaultdict
import warnings
warnings.filterwarnings('ignore')
ROOT = Path("/Users/manhnguyen/Project/NOC_DNA_V2")
RAW_DATA_DIR = ROOT / "data/PROVEDIt_1-5-Person CSVs UnFiltered"
OUTPUT_DIR = ROOT / "data/reconstructed"
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
def load_reference_genotypes():
"""Load reference genotypes from extracted data."""
print("📖 Loading reference genotypes...")
ref_df = pd.read_csv(OUTPUT_DIR / "reference_genotypes_raw.csv")
# Convert to nested dict: {study_id: {kit: {person_id: {marker: alleles}}}}
ref_db = defaultdict(lambda: defaultdict(lambda: defaultdict(dict)))
for _, row in ref_df.iterrows():
study = row['study_id']
kit = row['kit']
person = int(row['person_id'])
marker = row['marker']
alleles = str(row['alleles']).strip()
ref_db[study][kit][person][marker] = alleles
print(f"✅ Loaded reference genotypes for {len(ref_db)} studies")
for study in ref_db:
total_people = sum(len(ref_db[study][k]) for k in ref_db[study])
print(f" {study}: {total_people} people across {len(ref_db[study])} kits")
return ref_db
def extract_sample_genotype_v2(csv_path, num_contributors):
"""Extract observed genotypes from a sample CSV file - IMPROVED."""
try:
df = pd.read_csv(csv_path)
# Extract unique markers and their alleles
sample_genotype = {}
for _, row in df.iterrows():
marker = row['Marker']
# Get non-empty alleles (columns Allele 1-100)
alleles = []
for i in range(1, 101):
col_name = f'Allele {i}'
if col_name in df.columns:
allele = str(row[col_name]).strip()
# Filter out empty, NaN, OL (outlier), and invalid entries
if allele and allele != 'nan' and allele != '' and allele.upper() != 'OL':
try:
# Try to convert to float to validate it's a number
float(allele)
alleles.append(allele)
except:
# If not a number, still include it (could be AMEL format like X, Y)
if allele.upper() in ['X', 'Y', 'AM']:
alleles.append(allele.upper())
if alleles:
# Remove duplicates and sort for comparison
unique_alleles = sorted(set(alleles))
sample_genotype[marker] = ','.join(unique_alleles)
return sample_genotype
except Exception as e:
print(f" ⚠️ Error reading {csv_path.name}: {e}")
return {}
def normalize_allele(allele_str):
"""Normalize allele string for comparison."""
allele_str = str(allele_str).strip().upper()
# Handle AMEL specially
if allele_str in ['X', 'Y', 'AM']:
return allele_str
try:
return str(float(allele_str))
except:
return allele_str
def match_person_to_reference_v2(sample_markers, known_person_markers, verbose=False):
"""Calculate match score between sample and known person - IMPROVED."""
if not known_person_markers or not sample_markers:
return 0, 0, 0
# Find common markers
common_markers = set(sample_markers.keys()) & set(known_person_markers.keys())
if not common_markers:
return 0, 0, 0
# Count matching markers
matches = 0
perfect_matches = 0
for marker in common_markers:
# Normalize and parse alleles
sample_alleles = set([normalize_allele(a) for a in sample_markers[marker].split(',')])
ref_alleles = set([normalize_allele(a) for a in known_person_markers[marker].split(',')])
# Check if alleles match (at least one allele in common for heterozygotes)
# Or exact match for homozygotes
if sample_alleles & ref_alleles: # Any common allele
matches += 1
if sample_alleles == ref_alleles: # Exact match
perfect_matches += 1
match_score = matches / len(common_markers)
perfect_score = perfect_matches / len(common_markers)
return matches, match_score, perfect_score
def identify_known_unknown_contributors_v2(sample_genotype, ref_db, study_id, kit, num_contributors):
"""Identify which contributors in sample are known vs unknown - IMPROVED."""
known_matches = []
# Try to match against all known people in this study/kit
if study_id in ref_db and kit in ref_db[study_id]:
for person_id, person_markers in ref_db[study_id][kit].items():
matches, match_score, perfect_score = match_person_to_reference_v2(
sample_genotype, person_markers
)
# Use a threshold: need at least 70% matching alleles
if match_score >= 0.7:
known_matches.append((person_id, matches, match_score, perfect_score))
# Sort by perfect score (exact matches), then by match score
known_matches.sort(key=lambda x: (x[3], x[2]), reverse=True)
# For multi-person samples, deduplicate based on uniqueness
# If we have more matches than contributors, assume they're the same person identified multiple times
# by different markers
# Conservative approach: assume each high-quality match is one known person
unique_matches = []
used_people = set()
for person_id, matches, match_score, perfect_score in known_matches:
if person_id not in used_people and len(unique_matches) < num_contributors:
unique_matches.append((person_id, matches, match_score))
used_people.add(person_id)
num_known = len(unique_matches)
num_unknown = max(0, num_contributors - num_known)
return num_known, num_unknown, unique_matches
def process_all_samples_v2():
"""Process all raw sample CSV files and generate labels - IMPROVED."""
print("\n" + "=" * 80)
print("🔍 PHASE 2 v2: Parse Samples & Match to Reference (IMPROVED)")
print("=" * 80)
# Load reference
ref_db = load_reference_genotypes()
# Find all sample CSV files (exclude Known Genotypes files)
sample_files = []
for f in RAW_DATA_DIR.rglob("*.csv"):
if "Known Genotypes" not in f.name:
sample_files.append(f)
print(f"\n📂 Found {len(sample_files)} sample CSV files")
# Track results
results = []
stats = defaultdict(int)
# Process each file
for idx, csv_path in enumerate(sorted(sample_files)):
if idx % 100 == 0:
print(f" Processing {idx}/{len(sample_files)}...", end='\r')
# Extract folder info to determine number of contributors
path_parts = csv_path.parts
num_contributors = 1
study_id = "Unknown"
kit = "Unknown"
# Determine number of contributors from folder name
for part in path_parts:
if "1-Person" in part:
num_contributors = 1
elif "2-Person" in part:
num_contributors = 2
elif "3-Person" in part:
num_contributors = 3
elif "4-Person" in part:
num_contributors = 4
elif "5-Person" in part:
num_contributors = 5
# Extract study and kit from folder name
if "RD14" in part or "RD14-0003" in part:
study_id = "RD14-0003"
elif "RD12" in part or "RD12-0002" in part:
study_id = "RD12-0002"
if "IDPlus29" in part:
kit = "IDPlus29"
elif "IDPlus28" in part:
kit = "IDPlus28"
elif "GF29" in part:
kit = "GF29"
elif "F6C29" in part:
kit = "F6C29"
elif "PP16HS32" in part:
kit = "PP16HS32"
# Extract sample genotype
sample_genotype = extract_sample_genotype_v2(csv_path, num_contributors)
if not sample_genotype:
continue
# Match to reference
num_known, num_unknown, known_matches = identify_known_unknown_contributors_v2(
sample_genotype, ref_db, study_id, kit, num_contributors
)
# Validate
if num_known + num_unknown != num_contributors:
num_unknown = num_contributors - num_known
unknown_present = 1 if num_unknown > 0 else 0
results.append({
'sample_file': csv_path.name,
'study_id': study_id,
'kit': kit,
'num_contributors': num_contributors,
'num_known': num_known,
'num_unknown': num_unknown,
'unknown_present': unknown_present,
'num_markers': len(sample_genotype),
'top_match_score': known_matches[0][2] if known_matches else 0,
'matched_people': ';'.join([str(m[0]) for m in known_matches[:3]])
})
stats[f"{study_id}_{kit}_{num_contributors}P"] += 1
# Create results DataFrame
results_df = pd.DataFrame(results)
print(f"\n✅ Processed {len(results_df)} samples successfully")
print(f"\nDistribution by study/kit/contributors:")
for key in sorted(stats.keys()):
print(f" {key}: {stats[key]}")
# Check class balance
print(f"\nClass balance (unknown_present):")
print(f" No unknown (0): {(results_df['unknown_present'] == 0).sum()} ({100*(results_df['unknown_present'] == 0).sum()/len(results_df):.1f}%)")
print(f" Has unknown (1): {(results_df['unknown_present'] == 1).sum()} ({100*(results_df['unknown_present'] == 1).sum()/len(results_df):.1f}%)")
# Save results
output_file = OUTPUT_DIR / "sample_labels.csv"
results_df.to_csv(output_file, index=False)
print(f"\n💾 Saved: {output_file.name}")
return results_df
def main():
results_df = process_all_samples_v2()
print("\n" + "=" * 80)
print("✅ PHASE 2 v2 COMPLETE!")
print("=" * 80)
print(f"\nOutput summary:")
print(f" Total samples labeled: {len(results_df)}")
print(f" File: {OUTPUT_DIR / 'sample_labels.csv'}")
print(f"\nNext step: Phase 3 - Build final dataset with proper splits")
if __name__ == "__main__":
main()
|