annotate flexynesis_utils.py @ 5:466b593fd87e draft default tip

planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
author bgruening
date Fri, 04 Jul 2025 14:57:40 +0000
parents 525c661a7fdc
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
1 #!/usr/bin/env python
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
2
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
3 import argparse
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
4 import os
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
5 import sys
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
6 from pathlib import Path
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
7
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
8 import pandas as pd
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
9
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
10
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
11 def read_data(data_input, index=False):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
12 """Load CSV or TSV data file."""
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
13 try:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
14 file_ext = Path(data_input).suffix.lower()
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
15 sep = ',' if file_ext == '.csv' else '\t'
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
16 index_col = 0 if index else None
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
17
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
18 if file_ext in ['.csv', '.tsv', '.txt', '.tab', '.tabular']:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
19 return pd.read_csv(data_input, sep=sep, index_col=index_col)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
20 else:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
21 raise ValueError(f"Unsupported file extension: {file_ext}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
22 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
23 raise ValueError(f"Error loading data from {data_input}: {e}") from e
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
24
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
25
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
26 def binarize_mutations(df, gene_idx=1, sample_idx=2):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
27 """
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
28 Binarize mutation data by creating a matrix of gene x sample with 1/0 values.
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
29 """
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
30 # galaxy index is 1-based, convert to zero-based
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
31 gene_idx -= 1
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
32 sample_idx -= 1
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
33 # check idx
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
34 if gene_idx >= len(df.columns) or sample_idx >= len(df.columns):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
35 raise ValueError(f"Column indices out of bounds. DataFrame has {len(df.columns)} columns, "
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
36 f"but requested indices are {gene_idx} and {sample_idx}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
37 if gene_idx == sample_idx:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
38 raise ValueError("Gene and sample column indices must be different")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
39
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
40 # Get column names by index
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
41 gene_col = df.columns[gene_idx]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
42 print(f"Using gene column: {gene_col} (index {gene_idx})")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
43 sample_col = df.columns[sample_idx]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
44 print(f"Using sample column: {sample_col} (index {sample_idx})")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
45
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
46 # Check if columns contain data
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
47 if df[gene_col].isna().all():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
48 raise ValueError(f"Gene column (index {gene_idx}) contains only NaN values.")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
49 if df[sample_col].isna().all():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
50 raise ValueError(f"Sample column (index {sample_idx}) contains only NaN values.")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
51
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
52 # Group by gene and sample, count mutations
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
53 mutation_counts = df.groupby([gene_col, sample_col]).size().reset_index(name='count')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
54
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
55 # Create pivot table
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
56 mutation_matrix = mutation_counts.pivot(index=gene_col, columns=sample_col, values='count').fillna(0)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
57
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
58 # Binarize: convert any count > 0 to 1
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
59 mutation_matrix[mutation_matrix > 0] = 1
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
60
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
61 return mutation_matrix
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
62
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
63
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
64 def make_data_dict(clin_path, omics_paths):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
65 """Read clinical and omics data files into a dictionary."""
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
66 data = {}
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
67
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
68 # Read clinical data
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
69 print(f"Reading clinical data from {clin_path}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
70 try:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
71 clin = read_data(clin_path, index=True)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
72
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
73 if clin.empty:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
74 raise ValueError(f"Clinical file {clin_path} is empty")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
75 data['clin'] = clin
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
76 print(f"Loaded clinical data: {clin.shape[0]} samples, {clin.shape[1]} features")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
77 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
78 raise ValueError(f"Error reading clinical file {clin_path}: {e}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
79
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
80 # Read omics data
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
81 print(f"Reading omics data from {', '.join(omics_paths)}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
82 for path in omics_paths:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
83 try:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
84 name = os.path.splitext(os.path.basename(path))[0]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
85 df = read_data(path, index=True)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
86 if df.empty:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
87 print(f"Warning: Omics file {path} is empty, skipping")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
88 continue
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
89 data[name] = df
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
90 print(f"Loaded {name}: {df.shape[0]} features, {df.shape[1]} samples")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
91 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
92 print(f"Warning: Error reading omics file {path}: {e}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
93 continue
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
94
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
95 if len(data) == 1: # Only clinical data loaded
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
96 raise ValueError("No omics data was successfully loaded")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
97
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
98 return data
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
99
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
100
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
101 def validate_data_consistency(data):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
102 """Validate that clinical and omics data have consistent samples."""
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
103 clin_samples = set(data['clin'].index)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
104
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
105 for name, df in data.items():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
106 if name == 'clin':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
107 continue
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
108
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
109 omics_samples = set(df.columns)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
110
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
111 # Check for sample overlap
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
112 common_samples = clin_samples.intersection(omics_samples)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
113 if len(common_samples) == 0:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
114 raise ValueError(f"No common samples between clinical data and {name}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
115
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
116 missing_in_omics = clin_samples - omics_samples
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
117 missing_in_clin = omics_samples - clin_samples
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
118
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
119 if missing_in_omics:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
120 print(f"Warning: {len(missing_in_omics)} clinical samples not found in {name}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
121 if missing_in_clin:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
122 print(f"Warning: {len(missing_in_clin)} samples in {name} not found in clinical data")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
123
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
124 return True
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
125
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
126
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
127 def split_and_save_data(data, ratio=0.7, output_dir='.'):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
128 """Split data into train/test sets and save to files."""
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
129 # Validate data consistency first
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
130 validate_data_consistency(data)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
131
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
132 samples = data['clin'].index.tolist()
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
133
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
134 train_samples = list(pd.Series(samples).sample(frac=ratio, random_state=42))
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
135 test_samples = list(set(samples) - set(train_samples))
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
136
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
137 train_data = {}
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
138 test_data = {}
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
139
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
140 for key, df in data.items():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
141 try:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
142 if key == 'clin':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
143 train_data[key] = df.loc[df.index.intersection(train_samples)]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
144 test_data[key] = df.loc[df.index.intersection(test_samples)]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
145 else:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
146 train_data[key] = df.loc[:, df.columns.intersection(train_samples)]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
147 test_data[key] = df.loc[:, df.columns.intersection(test_samples)]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
148 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
149 print(f"Error splitting data {key}: {e}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
150 continue
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
151
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
152 # Create output directories
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
153 os.makedirs(os.path.join(output_dir, 'train'), exist_ok=True)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
154 os.makedirs(os.path.join(output_dir, 'test'), exist_ok=True)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
155
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
156 # Save train and test data
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
157 for key in data.keys():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
158 try:
5
466b593fd87e planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 3
diff changeset
159 train_data[key].to_csv(os.path.join(output_dir, 'train', f'{key}.tabular'), sep='\t')
466b593fd87e planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 3
diff changeset
160 test_data[key].to_csv(os.path.join(output_dir, 'test', f'{key}.tabular'), sep='\t')
3
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
161 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
162 print(f"Error saving {key}: {e}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
163 continue
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
164
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
165
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
166 def main():
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
167 parser = argparse.ArgumentParser(description='Flexynesis extra utilities')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
168
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
169 parser.add_argument("--util", type=str, required=True,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
170 choices=['split', 'binarize'],
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
171 help="Utility function: 'split' for spiting data to train and test, 'binarize' for creating a binarized matrix from a mutation data")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
172
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
173 # Arguments for split
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
174 parser.add_argument('--clin', required=False,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
175 help='Path to clinical data CSV file (samples in rows)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
176 parser.add_argument('--omics', required=False,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
177 help='Comma-separated list of omics CSV files (samples in columns)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
178 parser.add_argument('--split', type=float, default=0.7,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
179 help='Train split ratio (default: 0.7)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
180
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
181 # Arguments for binarize
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
182 parser.add_argument('--mutations', type=str, required=False,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
183 help='Path to mutation data CSV file (samples in rows, genes in columns)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
184 parser.add_argument('--gene_idx', type=int, default=0,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
185 help='Column index for genes in mutation data (default: 0)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
186 parser.add_argument('--sample_idx', type=int, default=1,
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
187 help='Column index for samples in mutation data (default: 1)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
188
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
189 # common arguments
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
190 parser.add_argument('--out', default='.',
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
191 help='Output directory (default: current directory)')
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
192
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
193 args = parser.parse_args()
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
194
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
195 try:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
196 # validate utility function
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
197 if not args.util:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
198 raise ValueError("Utility function must be specified")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
199 if args.util not in ['split', 'binarize']:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
200 raise ValueError(f"Invalid utility function: {args.util}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
201
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
202 if args.util == 'split':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
203 # Validate inputs
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
204 if not args.clin:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
205 raise ValueError("Clinical data file must be provided")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
206 if not args.omics:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
207 raise ValueError("At least one omics file must be provided")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
208 if not os.path.isfile(args.clin):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
209 raise FileNotFoundError(f"Clinical file not found: {args.clin}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
210 # Validate split ratio
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
211 if not 0 < args.split < 1:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
212 raise ValueError(f"Split ratio must be between 0 and 1, got {args.split}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
213
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
214 elif args.util == 'binarize':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
215 # Validate mutation data file
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
216 if not args.mutations:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
217 raise ValueError("Mutation data file must be provided")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
218 if not os.path.isfile(args.mutations):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
219 raise FileNotFoundError(f"Mutation data file not found: {args.mutations}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
220 # Validate gene and sample indices
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
221 if args.gene_idx < 0 or args.sample_idx < 0:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
222 raise ValueError("Gene and sample indices must be non-negative integers")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
223
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
224 # Create output directory if it doesn't exist
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
225 if not os.path.exists(args.out):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
226 os.makedirs(args.out)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
227
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
228 if args.util == 'split':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
229 # Parse omics files
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
230 omics_files = [f.strip() for f in args.omics.split(',') if f.strip()]
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
231 if not omics_files:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
232 raise ValueError("At least one omics file must be provided")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
233 # Check omics files exist
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
234 for f in omics_files:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
235 if not os.path.isfile(f):
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
236 raise FileNotFoundError(f"Omics file not found: {f}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
237 data = make_data_dict(args.clin, omics_files)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
238 split_and_save_data(data, ratio=args.split, output_dir=args.out)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
239
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
240 elif args.util == 'binarize':
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
241 mutations_df = read_data(args.mutations, index=False)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
242 if mutations_df.empty:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
243 raise ValueError("Mutation data file is empty")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
244
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
245 binarized_matrix = binarize_mutations(mutations_df, gene_idx=args.gene_idx, sample_idx=args.sample_idx)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
246 # Save binarized matrix
5
466b593fd87e planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 3
diff changeset
247 output_file = os.path.join(args.out, 'binarized_mutations.tabular')
466b593fd87e planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 3
diff changeset
248 binarized_matrix.to_csv(output_file, sep='\t')
3
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
249 print(f"Binarized mutation matrix saved to {output_file}")
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
250
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
251 except Exception as e:
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
252 print(f"Error: {e}", file=sys.stderr)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
253 sys.exit(1)
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
254
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
255
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
256 if __name__ == "__main__":
525c661a7fdc planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
257 main()