from matplotlib.pyplot import *
from numpy import average
from scipy import interpolate
from scipy.stats import pearsonr, spearmanr, sem, ks_2samp,wilcoxon,mannwhitneyu
import pandas as pd

protein_data = pd.read_csv('martin2012_ostreo_diurnal_N15_proteomics.csv')
timepoints = protein_data.columns[5:]
nC = len(protein_data.columns)

#estimate achievable maximum fraction of labelling (with a small additional buffer)
max_label_fraction = max(protein_data[timepoints[-1]])+0.01

rescaled_data = pd.DataFrame(columns=timepoints,index=protein_data.index)
rescaled_data[timepoints] = np.log((max_label_fraction-protein_data[timepoints])/max_label_fraction)

timepoint_ints = [int(x) for x in timepoints]

protein_data['Deg rate (h-1)']= rescaled_data.apply(lambda x : -np.polyfit(timepoint_ints,x[timepoints],1)[0],axis=1)
protein_data['Half life (h)'] = np.log(2)/protein_data['Deg rate (h-1)']
protein_data['Half life (days)'] = protein_data['Half life (h)']/24

protein_data.drop('Deg rate',inplace=True,axis=1)

protein_data = protein_data[['Ot gene Idb','Deg rate (h-1)','Half life (h)']]
protein_data.to_csv('Otauri_degradation_rates.csv',index=None)