| |
| |
| import os |
| import numpy as np |
| import pandas as pd |
| from collections import Counter |
| from tqdm import tqdm |
|
|
|
|
| def pfam_level_metadata(seed_alignment_dir: str, |
| pfam: str): |
| """ |
| get pfam-level metadata, including: |
| - clan: clan that pfam belongs in (if any) |
| - type: type of pfam (domain, family, etc) |
| - width: width of the MSA |
| - depth: how many sequences per pfam |
| - percent_gaps: out of every charcter in the MSA, how many |
| are gap chars? (NOT normalized by any particular |
| sequence length) |
| |
| inputs: |
| ------- |
| - seed_alignment_dir: where pfam MSA seed alignments are |
| - pfam: name of the PFam (PF#####) |
| |
| returns: |
| -------- |
| - out_dict: PFam-level metadata |
| """ |
| |
| out_dict = {'pfam': pfam} |
| msa_width_found = False |
| seq_count = 0 |
| char_count = 0 |
| gaps_count = 0 |
| with open(f'{seed_alignment_dir}/{pfam}.seed', 'r') as f: |
| for line in f: |
| if line.startswith('#=GF CL'): |
| clan_name = line.strip().split()[-1] |
| out_dict['clan_name'] = clan_name |
| |
| elif line.startswith('#=GF TP'): |
| pfam_type = line.strip().split()[-1] |
| out_dict['type'] = pfam_type |
| |
| elif not line.startswith('#'): |
| gapped_seq = line.strip().split()[-1] |
| gapped_seq = gapped_seq.upper() |
| |
| |
| seq_count += 1 |
| |
| |
| if not msa_width_found: |
| out_dict['msa_width'] = len(gapped_seq) |
| msa_width_found = True |
| |
| |
| char_count += len(gapped_seq) |
| gaps_count += gapped_seq.count('.') |
| |
| out_dict['msa_depth'] = seq_count |
| out_dict['percent_gaps'] = gaps_count / char_count |
| out_dict['clan_name'] = out_dict.get('clan_name','') |
| out_dict['type'] = out_dict.get('type','') |
| |
| return out_dict |
|
|
|
|
| def serially_pfam_level_metadata(pfam_seed_file: str, |
| seed_alignment_dir: str): |
| """ |
| use pfam_level_metadata() on all files in a folder |
| |
| inputs: |
| ------- |
| - pfam_seed_file: the original seed file (for figuring out prefix) |
| - seed_alignment_dir: where seed files are located |
| |
| returns: |
| -------- |
| - all_meta: the dataframe of stats |
| |
| outputs: |
| -------- |
| - pfam_level_metadata_file: f'{prefix}_PFAM-METADATA.tsv |
| """ |
| |
| |
| |
| pfam_lst = [file.replace('.seed','') for file in os.listdir(seed_alignment_dir) |
| if file.startswith('PF') and file.endswith('.seed')] |
|
|
| all_meta = [] |
| for pfam in tqdm(pfam_lst): |
| out_dict = pfam_level_metadata(seed_alignment_dir = seed_alignment_dir, |
| pfam = pfam) |
| all_meta.append(out_dict) |
| del out_dict |
| |
| all_meta = pd.DataFrame(all_meta) |
| return all_meta |
|
|
|
|
| def clan_level_metadata(pfam_seed_file: str): |
| """ |
| get clan-level metadata, including: |
| - total pfams |
| - total sequences |
| - list of pfams in each clan |
| |
| inputs: |
| ------- |
| - pfam_seed_file: the original seed file (for figuring out prefix) |
| |
| returns: |
| -------- |
| - clan_metadata: the dataframe of stats |
| |
| outputs: |
| -------- |
| - clan_level_metadata_file: f'{prefix}_CLAN-METADATA.tsv |
| |
| """ |
| prefix = pfam_seed_file.split('.')[0] |
| pfam_level_metadata_file = f'{prefix}_PFAM-METADATA.tsv' |
| |
| |
| path = "/".join( prefix.split('/')[:-1] ) |
| pfam_level_metadata_file_without_path = pfam_level_metadata_file.split('/')[-1] |
| err = f'{pfam_level_metadata_file} not found!' |
| assert pfam_level_metadata_file_without_path in os.listdir(path), err |
| del path, pfam_level_metadata_file_without_path |
| |
| df = pd.read_csv(pfam_level_metadata_file, sep='\t', index_col = 0) |
| df = df.fillna('') |
| clan_counts = dict(Counter(df['clan_name'])) |
| if '' in clan_counts.keys(): |
| del clan_counts[''] |
| |
| |
| clan_metadata = [] |
| for clan in clan_counts.keys(): |
| sub_df = df[df['clan_name'] == clan] |
| num_seqs = sub_df['msa_depth'].sum() |
| num_pfams = len(sub_df) |
| |
| out_dict = {'clan_name': clan, |
| 'num_pfams': num_pfams, |
| 'num_seqs': num_seqs, |
| 'pfams': '; '.join( sub_df['pfam'].tolist() ) |
| } |
| clan_metadata.append(out_dict) |
| |
| clan_metadata = pd.DataFrame(clan_metadata) |
| return clan_metadata |
|
|
|
|
