Mercurial > repos > estrain > lissero

comparison variant4b.py @ 0:bc556481a1fb draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit e12eb56d1744da8f7af8ca1819e2617c83fb17a8
author estrain
date Fri, 13 Mar 2026 12:01:17 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:bc556481a1fb
1 import sys
2
3 def identify_variants_with_genes(input_file_path, output_file_path, simple_name):
4 # Define the genes of interest
5 genes_of_interest = ['LMO0737', 'ORF2110', 'ORF2819']
6
7 # Open the input file and read its lines
8 with open(input_file_path, 'r') as file:
9 lines = file.readlines()
10
11 # Check if the file has more than just the header
12 if len(lines) <= 1:
13 print("Input file does not contain enough data.")
14 return
15
16 # Extract the column headers and find the indices of the genes of interest
17 headers = lines[0].strip().split('\t')
18 gene_indices = [headers.index(gene) for gene in genes_of_interest]
19 serotype_index = headers.index('SEROTYPE')
20
21 # Modify the header to include the new first column
22 modified_header = f"FileName\t{lines[0]}"
23
24 # Initialize a list to hold the modified lines, starting with the modified header
25 modified_lines = [modified_header]
26
27 # Process each data line in the input file
28 for line in lines[1:]:
29 data = line.strip().split('\t')
30 # Check if the genes of interest are all present (marked as "FULL")
31 if all(data[index] == 'FULL' for index in gene_indices):
32 # Modify the SEROTYPE column to "4b variant"
33 data[serotype_index] = "4b variant"
34 # Prepend the simple name to the line
35 modified_line = f"{simple_name}\t" + '\t'.join(data) + '\n'
36 # Add the modified line to the list
37 modified_lines.append(modified_line)
38
39 # Write the modified lines to the output file
40 with open(output_file_path, 'w') as file:
41 file.writelines(modified_lines)
42
43 print(f'Results written to {output_file_path}')
44
45 if __name__ == "__main__":
46 if len(sys.argv) != 4:
47 print("Usage: python script.py <input_file_path> <output_file_path> <simple_name>")
48 sys.exit(1)
49
50 input_file_path = sys.argv[1]
51 output_file_path = sys.argv[2]
52 simple_name = sys.argv[3]
53 identify_variants_with_genes(input_file_path, output_file_path, simple_name)
54