Title: Python scripts for in silico evolution of codon usage 

 
random_codon.py
Evolution of codon usage using known mutation frequencies with random direction of mutation

#####################################################################################

#!/usr/bin/python
"""
Global patterns of genome evolution associated with generalist parasitism
Thomas Badet, Remi Peyraud, Malick Mbengue, Olivier Navaud, Mark Derbyshire, Richard P. Oliver, Adelin Barbacci, Sylvain Raffaele
"""

from pandas import *
import random as rd

def read(path):
	data = read_csv(path)
	return data

def get_target(data):
	# Generates random mutation
	N_mut = rd.randint(1,len(data))
	t = rd.sample(data.Codon.values,rd.randint(1,len(data)))
	m = []
	for ii in range(0,len(t)):
		m.append(rd.random()) 
	# Normalizes array of mutation
	m = np.array(m)/sum(m)
	return t,m.tolist()


def mutation(data):
	Datab = data.copy()
	for i,row in data.iterrows():
		# Synonymous mutations
		syno = data[(row.AA==data.AA) & (data.Codon != row.Codon)]
		if(len(syno)!=0):
			# chosing randomly mutations
			target_s,freq = get_target(syno)
			# Numbers of codons added to every mutations
			num = (np.array(freq)*row.Syn_mutation_rate*row.Number*0.01).tolist()
			for c,n1 in zip(target_s,num):
				Datab.Number[c==Datab.Codon] += n1
			# Updating source codons
			Datab.Number[row.Codon==Datab.Codon] -= sum(num)
		# Not synonymous mutations
		n_syno = data[data.AA != row.AA]
		target_ns,freq = get_target(n_syno)
		# Numbers of codons added to every mutations
		num = (np.array(freq)*row.ns_mutation_rate*row.Number*0.01).tolist()
		for c,n1 in zip(target_ns,num):
			Datab.Number[c==Datab.Codon] += n1
		# Updating source codons
		Datab.Number[row.Codon==Datab.Codon] -= sum(num)
	return Datab

def get_change(data,tps,iterat):
	# Computes the number of optimal, not-optimal and undifferenciated codons
	opt = float(sum(data.Number[data.optimal=="yes"]))/float(sum(data.Number))
	n_opt = float(sum(data.Number[data.optimal=="no"]))/float(sum(data.Number))
	ind = float(sum(data.Number[data.optimal=="diff"]))/float(sum(data.Number))
	d = DataFrame([{'iteration':iterat,'t':tps,'opti':opt,'non_opti':n_opt,'diff':ind},])
	return d

if __name__=="__main__":
	# Input file
	D = read_csv('./sclero.txt',sep='\t',usecols=['Codon','AA','Number','Syn_mutation_rate','ns_mutation_rate','optimal'])
	# Number of repetitions of simulations
	for kk in range(0,10):
		if kk == 0:
			# output table
			res = get_change(D,0,kk)
		else :
			res = concat([res,get_change(D,0,kk)])

		D1 = D.copy()
		for ii in range(1,1000):
			# Simulation of long time evolution
			print("Iteration "+str(kk)+" time "+str(ii))
			D1 = mutation(D1)
			res = concat([res,get_change(D1,ii,kk)])
	# writing results
	res.to_csv("./res_python_sclero_ok_3.csv",sep='\t')

#####################################################################################

non_random_codon.py
Evolution of codon usage using fixed mutation frequencies and fixed substitution patterns

#####################################################################################

#!/usr/bin/python

"""
Global patterns of genome evolution associated with generalist parasitism
Thomas Badet, Remi Peyraud, Malick Mbengue, Olivier Navaud, Mark Derbyshire, Richard P. Oliver, Adelin Barbacci, Sylvain Raffaele
"""

from pandas import *


def read(path):
	data = read_csv(path)
	return data

def mutation(data):
	# Simulates mutations 
	Datab = data.copy()
	for i,row in data.iterrows():
		mut = row[3:]*row[0]*0.01
		for codon in mut.index:
			Datab.N[Datab.Reference==codon]+=mut[codon]
		Datab.N[Datab.Reference == row.Reference] -= sum(mut)
	return Datab

def get_change(data,tps):
	# Returns the numbers of optimal, not-optimal and undiffercentiated codons
	opt = float(sum(data.N[data.optimal=="yes"]))/float(sum(data.N))
	n_opt = float(sum(data.N[data.optimal=="no"]))/float(sum(data.N))
	ind = float(sum(data.N[data.optimal=="diff"]))/float(sum(data.N))
	d = DataFrame([{'t':tps,'opti':opt,'non_opti':n_opt,'diff':ind},])
	return d

if __name__=="__main__":
	# Input file 
	D1 = read_csv('./Codon_mutations_Zymo_3.csv',sep='\t')
	# Number of optimal, not optimal and not differcentiated codons
	res = get_change(D1,0)
	for ii in range(1,1000):
		# Computing mutation
		D1 = mutation(D1)
		# Creating table of results
		res = concat([res,get_change(D1,ii)])
	# Storing output file
	res.to_csv("./res_python_Zymo_det.csv",sep='\t')
	print("finished")