# Copyright (c) 2025 Isom Lab # Released under the MIT License; see LICENSE at the repository root. # Specific UniProt collection... ################################################################################################# # hit_file_path = "/projectnb/isomlab/dan/in-and-out/output/alphaFold_uniprot_v4/2023.11.12.7tmps/2023.11.12.7tmps-analysis_final/" # hit_file_name = "2023.11.12.7tmps_results.csv" hit_file_path = "/projectnb/isomlab/dan/in-and-out/output/alphaFold_uniprot_v4/2025.09.27.7tmps-foldseek_e0p01/2025.09.27.7tmps-foldseek_e0p01-analysis_foldseek_e0p01/" hit_file_name = "2025.09.27.7tmps-foldseek_e0p01_results.csv" # hit_file_path = "/projectnb/isomlab/dan/in-and-out/output/alphaFold_uniprot_v4/2025.08.20.7tmp_sensitivity_test/mglu/2025.08.20.7tmps-rhod_mglu-analysis_test/" # hit_file_name = "2025.08.20.7tmps-rhod_mglu_results.csv" # Calculation details... ################################################################################################# cluster = "triton" number_of_calcs_per_job = 100 # Takes about 4 minutes per 500 jobs.. superdark_script = "/home/dgi5/scripts/proteomes-214M-v4/1-job-code/1-query/3-alignResults.py" # UM cluster information... # Pegasus queues: debug or general # Triton queues: short or normal queue = "normal" # Time to sleep in seconds between each check for running job(s)... sleep_wait = 30 """ ========================================================================================================= # Script: **Batch TM-align Structural Alignment** # Description: # This script automates the **pairwise structural alignment** of query structures using **TM-align** # against a set of **UniProt-derived protein structures**. It distributes the calculations across # a **high-performance computing cluster (Triton)**, monitors job execution, logs results, and # performs cleanup operations. ========================================================================================================= 📌 **CONFIGURATION PARAMETERS** --------------------------------------------------------------------------------------------------------- 🔹 **File Paths** - `hit_file_path` → Directory containing the list of target structures for alignment. - `hit_file_name` → Name of the CSV file with UniProt IDs of structures to be aligned. 🔹 **Cluster Configuration** - `cluster` → Specifies the cluster (**Triton**). - `queue` → Specifies the job queue (**normal**). - `number_of_calcs_per_job` → Controls how many calculations are submitted per job. - `superdark_script` → The Python script (`3-alignResults.py`) that processes the alignments. 🔹 **Job Monitoring** - `sleep_wait` → Defines the waiting time (seconds) between job status checks. ========================================================================================================= 📌 **DEPENDENCIES** --------------------------------------------------------------------------------------------------------- - `os`, `shutil`, `subprocess`, `time` → Used for file handling, system commands, and cluster job management. ========================================================================================================= 📌 **CLASSES** --------------------------------------------------------------------------------------------------------- 🔹 **`Result`** - Represents a **pairwise alignment result**. - **Attributes:** - `tm_score`: **Template modeling score** (higher = better match). - `rmsd`: **Root Mean Square Deviation** (lower = better match). - `query_length`: Length of the **query** structure. - `n_align`: Number of **aligned residues**. - `query_pdb_file_path`: File path for the **query** structure. - `target_pdb_file_path`: File path for the **target** structure. - **Methods:** - `__repr__()`: Returns a **CSV-formatted** string of the alignment result. ========================================================================================================= 📌 **EXECUTION FLOW** --------------------------------------------------------------------------------------------------------- 🔹 **Step 1: Load and Parse Structural Matches** - Reads **target structures** from the CSV file (`hit_file_name`). - Creates a list of **`Result`** objects. 🔹 **Step 2: Split Targets into Job Groups** - Divides the list into **evenly-sized** groups. - Ensures each **job processes `number_of_calcs_per_job` structures**. 🔹 **Step 3: Create Job Submission Directories** - Creates **three directories**: 1️⃣ **`input_files_directory/`** → Stores job input files. 2️⃣ **`save_directory/`** → Stores alignment results. 3️⃣ **`log_directory/`** → Stores job logs. 🔹 **Step 4: Generate Job Submission Files** - Writes the **pairwise structure pairs** into job-specific input files. - Prepares the **input file paths** for batch submission. 🔹 **Step 5: Submit Jobs to the Cluster** - **Uses `bsub`** to submit each job to the Triton **batch queue**. - Assigns: ✅ Job name → `"j3-"`. ✅ Standard output/error logs. 🔹 **Step 6: Monitor Job Completion** - Uses **`bjobs`** to check for running jobs. - **Waits (`sleep_wait` seconds) before rechecking.** 🔹 **Step 7: Cleanup** - **Deletes intermediate rotation matrix files** (`.txt`) generated by TM-align. ========================================================================================================= 📌 **SUMMARY** --------------------------------------------------------------------------------------------------------- ✅ Automates large-scale **protein structure alignment**. ✅ Runs **distributed jobs** on the **Triton cluster**. ✅ Implements **job monitoring & error handling**. ✅ Cleans up **temporary calculation files** after completion. ========================================================================================================= """ # Dependencies... ################################################################################################# from os import mkdir, listdir, remove from os.path import exists from shutil import rmtree from subprocess import run from time import sleep # Classes... ################################################################################################# class Result: def __init__(self, result_file_line): self.result_file_line = result_file_line self.data = result_file_line.strip().split(",") self.tm_score = float(self.data[0]) # TM score normalized by the target structure self.rmsd = float(self.data[1]) self.query_length = int(self.data[2]) self.n_align = int(self.data[3]) self.query_pdb_file_path = self.data[4] self.target_pdb_file_path = self.data[5] def __repr__(self): return_string = "%f,%f,%i,%i,%s,%s\n" % ( self.tm_score, self.rmsd, self.query_length, self.n_align, self.query_pdb_file_path, self.target_pdb_file_path ) return return_string ################################################################################################# # # Main code... # ################################################################################################# # Get each initial structure similarity match... ################################################################################################## result_objects = [] f = open(hit_file_path + hit_file_name, "r") for line in f: try: result_object = Result(line) result_objects.append(result_object) except: pass f.close() # Evenly the split matches... ################################################################################################## myList = result_objects N = int(len(myList)/number_of_calcs_per_job) if not N: N = 1 result_groups = [myList[(i*len(myList))//N:((i+1)*len(myList))//N] for i in range(N)] # Make a directory for the job input files... ################################################################################################## input_files_directory = hit_file_path + hit_file_name.split(".csv")[0] + "_input_files/" if input_files_directory != "/": if not exists(input_files_directory): mkdir(input_files_directory) else: rmtree(input_files_directory) mkdir(input_files_directory) # Make a save directory... ################################################################################################## save_directory = hit_file_path + hit_file_name.split(".csv")[0] + "-unfiltered_hits/" if save_directory != "/": if not exists(save_directory): mkdir(save_directory) else: rmtree(save_directory) mkdir(save_directory) # Make a log directory... ################################################################################################## log_directory = hit_file_path + hit_file_name.split(".csv")[0] + "-unfiltered_hits_logs/" if log_directory != "/": if not exists(log_directory): mkdir(log_directory) else: rmtree(log_directory) mkdir(log_directory) # Write initial match groups to job submission input files... ################################################################################################## job_number, result_object_file_paths = 1, [] for result_group in result_groups: if not result_group: continue result_object_group_string = "" for result_object in result_group: result_object_group_string += str(result_object) f = open(input_files_directory + str(job_number) + "_input_file.txt", "w") f.write(result_object_group_string) f.close() # Save the input file path for submission... result_object_file_paths.append(input_files_directory + str(job_number) + "_input_file.txt") job_number += 1 # Farm the job submissions... ################################################################################################################ job_number = 1 for result_object_file_path in result_object_file_paths: i_log_directory = log_directory + "job-"+str(job_number)+".log" i_error_directory = log_directory + "job-"+str(job_number)+"-error.log" run(["bsub", "-q", queue, "-J", "j3-" + str(job_number), "-o", i_log_directory, "-e", i_error_directory, "python", superdark_script, "-f", result_object_file_path, "-c", cluster, "-s", save_directory]) job_number += 1 # Monitor the jobs and delete intermediate, TMAlign-generated .txt rotation matrix files... ################################################################################################################ jobs_running = 1 while jobs_running: t = run(["bjobs"], capture_output=True) stdout = t.stdout.decode() if "j3-" in stdout: print("still running...3-alignResults-jobParser.py") sleep(sleep_wait) continue else: jobs_running = 0 # Remove temporary mobile files... print("Removing temporary calculation files....") filenames = listdir(save_directory) for filename in filenames: for phrase in [".txt"]: if phrase in filename: print(filename) remove(save_directory + filename) break