EGF Stimulation of 184A1 Cells (Wolf Yadlin, 2007)

#Supporting packages for analysis
import numpy as np
import pandas as pd

#KSTAR imports
from kstar import config, helpers, calculate
from kstar.plot import DotPlot


#Set matplotlib defaults for arial 12 point font
from matplotlib import rcParams
rcParams['font.family'] = 'sans-serif'
rcParams['font.sans-serif'] = ['Arial']
rcParams['font.size'] = 12
import matplotlib.pyplot as plt

#where supplementary data was downloaded to
SUPPLEMENTS_DIR = './'

#Directory where KSTAR Supplemental data was set (From https://figshare.com/articles/dataset/KSTAR_Supplementary_Data/14919726)
odir = SUPPLEMENTS_DIR+'Supplements/SupplementaryData/Control_Experiments/EGF_184A1_WolfYadlin2007/'

#load the Mann Whitney activities and FPR for Tyrosine predictions,
#it will be faster and less data than loading all KSTAR outputs
activities = pd.read_csv(odir+'/RESULTS/EGF_HMEC_Y_mann_whitney_activities.tsv', sep='\t', index_col=0)
fpr = pd.read_csv(odir+'/RESULTS/EGF_HMEC_Y_mann_whitney_fpr.tsv', sep='\t', index_col=0)



#load kinase map from supplementary data
KINASE_MAP =  pd.read_csv(SUPPLEMENTS_DIR+'SupplementaryData/Map/globalKinaseMap.csv', index_col = 0)
#set preferred kinase names from the kinase map (make a kinase_dict)
kinase_dict = {}
for kinase in activities.index:
    kinase_dict[kinase] = KINASE_MAP.loc[kinase,'Preferred Name']

Agglomerative clustering of kinases

results = activities
results = -np.log10(results)


fig, axes = plt.subplots(figsize = (9,12),
        nrows = 1, ncols = 2,
        sharex = 'col',
        sharey = 'row',
        gridspec_kw = {
            'width_ratios':[0.1,1]
        },)
fig.subplots_adjust(wspace=0, hspace=0)

dots = DotPlot(results,
               fpr,
               figsize = (9,12),
               dotsize = 10,
               legend_title='-log10(p-value)',
               kinase_dict=kinase_dict)


dots.cluster(orientation = 'left', ax = axes[0], method='ward')

dots.dotplot(ax = axes[1])
plt.xlabel('Time', FontSize=12)
plt.xticks(rotation = 45, FontSize=12)
plt.yticks(FontSize=12)
plt.savefig(odir+'EGF_HMEC_all.pdf', bbox_inches='tight')

Agglomerative clustering of only significant kinases

results = activities
results = -np.log10(results)

fig, axes = plt.subplots(figsize = (9,12),
        nrows = 1, ncols = 2,
        sharex = 'col',
        sharey = 'row',
        gridspec_kw = {
            'width_ratios':[0.1,1]
        },)
fig.subplots_adjust(wspace=0, hspace=0)

dots = DotPlot(results,
               fpr,
               figsize = (9,12),
               dotsize = 10,
               legend_title='-log10(p-value)',
               kinase_dict=kinase_dict)

dots.drop_kinases_with_no_significance()

dots.cluster(orientation = 'left', ax = axes[0], method='ward')

dots.dotplot(ax = axes[1])
plt.xlabel('Time', FontSize=12)
plt.xticks(rotation = 45, FontSize=12)
plt.yticks(FontSize=12)
plt.savefig(odir+'EGF_HMEC_sigKinases.pdf', bbox_inches='tight')