GAPO is a in silico genetic algorithm used to optimize proteins for a desired function, such as stability and affinity. The algorithm mimetizes the evolutionary process, by recombining and adding mutations to the best sequences in order to generate a new population with higher diversity and optimized for the given objetive function.
git clone https://github.com/izzetbiophysicist/prot_eng_GA.git
First of all, you must download PyRosetta.
PyRosetta4 Download
Additional help for downloading and installing and PyRosetta (source:Sari Sabban youtube channel )
conda create --name GA_env --file requirements.txt
tar -xvf PyRosetta4.Release.python310.linux.release-370.tar.bz2
cd PyRosetta4.Release.python310.linux.release-370/setup
conda activate GA_env
python setup.py install
https://github.com/chavesejf/pbee.
After install pbee and all dependencies, remember to put in both pbee and GA folders and files on the same directory.
genetic_algorithm.py = contains main function
apt_functions.py = contains a collection of objective functions for the genetic algorithm
GAprot.py imports genetic_algorithmpy and apt_functions.py, configures initial population and carries out the optimization
| Parameter | Description |
|---|---|
| opt_direction | Selects if the objective function goes up or down |
| apt_function | Select between "apt_thread" for pbee and "apt_rosetta" for rosetta. |
| gene_values | Values that vectors can assume |
| gene_type | Selects between discrete or continuous genes |
| vector_size | Size of the vectors |
| selection_methods | Elitist or tournament selection |
| threahds | Parallel processing (not supported) |
| n_survivor | Number of survivors in each selection |
| tournament_size | Number of individuals selected in each tournament |
| lista_fixed | List of residues to maintain during evolution |
| initial_population | An initial population can be given. Otherwise, a random population is created |
| file_name | Output log file, Sequences - dG - population |
| dg_method | Select between "fold" and "bind". |
| cpus | Numbers of CPU usage to paralelize |
Follow below a usage example of the function used to optimize the CDRs from a scFv complexed to CD19 (script GAprot.py)
####### Loading libraries
import os
from genetic_algorithm_rosetta import genetic_algo
from apt_function import *
import apt_function
from pyrosetta import *
from rosetta.core.pack.task import TaskFactory
from rosetta.core.pack.task import operation
import numpy as np
from numpy.random import uniform
from random import sample
# Initialize PyRosetta
pyrosetta.init()
# Create a scoring function using the "ref2015_cart.wts" weight set
scorefxn = pyrosetta.create_score_function("ref2015_cart.wts")
# Creates pose object from input PDB
starting_pose = pose_from_pdb('CD19_scFv_relax.pdb')
# Relax the starting pose by packing and relaxing it iteratively for 3 times
starting_pose = apt_function.pack_relax(starting_pose = starting_pose, scorefxn = scorefxn, times_to_relax = 3)
scorefxn(starting_pose)
# Define a list of single-letter amino acid codes
gene_values = ['A','C','D','E','F','G','H','I','K','L','M','N','P','Q','R','S','T','V','W','Y']
starting_pose_seq = [x for x in starting_pose.sequence()]
# Residues to mutate during optimization
CDRs = [62,63,64,65,66,67,68,69,70,71,72,88,89,90,91,92,93,94,127,128,129,130,131,132,133,134,135,186,187,188,189,190,191,192,212,213,214,215,216,257,258,259,260,261,262,263,264,265,266,267,268,269]# Chain identifier for the fixed residues
# Generate an initial population of protein structures for optimization and return the list of residues to maintain locked during the evolution
# Pop size receives the size of the population to be generated
# fixed_residues_list receives the list of residues to mutate during the evolution
# chains receive a list of the chains from the respective residues to mutate
init_population, list_fixed_index = apt_function.Generate_random_population(starting_pose = starting_pose,
pop_size = 50,
fixed_residues_list = CDRs,
chains = ["C", "D"])
GA = genetic_algo(pose=starting_pose, opt_direction='down',initial_population = init_population, gene_values=gene_values, gene_type='discrete',
vector_size=len(starting_pose_seq), pop_size=len(init_population), mutation_rate=0.025, segment_fluctuation=0,
apt_function=apt, selection_method='tournament', threads=False,
convergence_threshold=0, n_cycles=4, tournament_cycles=int(np.round(len(init_population)/4)), tournament_size=4, benchmark=False,
lista_fixed=list_fixed_index, crossing_over_type='mask', file_name="teste_1.txt", dg_method="bind", cpus = 5)
GA.execute()