def get_tukey(exp, df_all, measure):
# Tukey posthoc analysis
# See https://jpktd.blogspot.com/2013/03/multiple-comparison-and-tukey-hsd-or_25.html
# And https://code.google.com/archive/p/qsturng-py/
# And https://stackoverflow.com/questions/48200699/how-can-i-get-p-values-of-each-group-comparison-when-applying-the-tukey-s-hones
# q, res_table, std_pairs, etc can be found from print(dir(result)) which will list all possible calculations
if len(df_all.groupby('strain').count()) >= 3:
df_tukey = df_all[np.isfinite(df_all[measure])]
mc = MultiComparison(df_tukey[measure], df_tukey['strain'])
result = mc.tukeyhsd()
p = psturng(np.abs(result.meandiffs / result.std_pairs),
len(result.groupsunique), result.df_total)
df_pairs = pd.DataFrame({
'group1': [
result._results_table[1][0], result._results_table[2][0],
result._results_table[3][0]
],
'group2': [
result._results_table[1][1], result._results_table[2][1],
result._results_table[3][1]
],
'p_value':
[np.around(p[0], 4),
np.around(p[1], 4),
np.around(p[2], 4)]
})
else:
df_pairs = pd.DataFrame({'group1': [], 'group2': [], 'p_value': []})
file_out = exp.name + '_coupling_' + measure + '_' + '_tukey_' + '.csv'
pfiles.save_csv(df_pairs, file_out, exp.dir_tukey, False)
return df_pairs
def get_hyp_coupling(exp, df_all, dir_in):
os.chdir(dir_in)
list_groups = list() # creates list of group dfs for attempted
list_groups = pfiles.get_groups(
exp, df_all) # Fills list of group dfs for attempted
# If strain = na then drop
df_anova = pd.DataFrame()
df_tukey = pd.DataFrame()
for meas in exp.measures:
df_new_anova = pd.DataFrame()
df_new_tukey = pd.DataFrame()
new_anova = get_anova(exp, list_groups, meas)
new_anova['measure'] = meas
new_anova.set_index('measure', inplace=True)
df_anova = df_anova.append(new_anova)
new_tukey = get_tukey(exp, df_all, meas)
new_tukey['measure'] = meas
new_tukey.set_index('measure', inplace=True)
df_tukey = df_tukey.append(new_tukey)
file_anova = exp.name + '_coupling_all_anovas.csv'
file_tukey = exp.name + '_coupling_all_tukey.csv'
pfiles.save_csv(df_anova, file_anova, dir_in + exp.dir_anova, True)
pfiles.save_csv(df_tukey, file_tukey, dir_in + exp.dir_tukey, True)
def sort_connected(data, exp):
data_connected = data.copy()
for index, row in data_connected.iterrows():
if row['cc'] < exp.connection:
data_connected.drop(index, inplace=True)
file_out = exp.name + '_coupling_connected.csv'
pfiles.save_csv(data_connected, file_out, exp.dir_in, True)
return data_connected, file_out
def sort_nmda50(data, exp):
data_nmda50 = data.copy()
for index, row in data_nmda50.iterrows():
if row['conc_nmda'] != 50:
data_nmda50.drop(index, inplace=True)
file_out = exp.name + '_coupling_nmda50.csv'
pfiles.save_csv(data_nmda50, file_out, exp.dir_in, True)
return data_nmda50, file_out
def get_filtered(data, exp):
# Filter data based on Vm and Rinput
# filter [0] = vm1
# filter [1] = rin1
data = data.drop(data[data.vm1 > exp.filters[0]].index)
data = data.drop(data[np.isnan(data.vstim1) == True].index)
file_out = exp.name + '_coupling_attempted.csv'
pfiles.save_csv(data, file_out, exp.dir_in, True)
return data, file_out
def sort_NMDA(data, exp):
data_NMDA = data.copy()
for index, row in data_NMDA.iterrows():
if row['cc'] < 0.1 and row['cc_nmda'] > 0.1 and row['cc_pchange'] > 0:
pass
else:
data_NMDA.drop(index, inplace=True)
file_out = exp.name + '_coupling_created.csv'
pfiles.save_csv(data_NMDA, file_out, exp.dir_in, True)
return data_NMDA, file_out
def sort_weakened(data, exp):
data_weakened = data.copy()
for index, row in data_weakened.iterrows():
if row['cc'] >= 0.1 and row['cc_pchange'] < 0:
pass
else:
data_weakened.drop(index, inplace=True)
file_out = exp.name + '_coupling_weakened.csv'
pfiles.save_csv(data_weakened, file_out, exp.dir_in, True)
return data_weakened, file_out
def sort_weak(data, exp):
data_weak = data.copy()
for index, row in data_weak.iterrows():
if row['cc'] < exp.strong and row['cc'] >= 0.1 and row['cc_nmda'] >= 0:
pass
else:
data_weak.drop(index, inplace=True)
file_out = exp.name + '_coupling_weak.csv'
pfiles.save_csv(data_weak, file_out, exp.dir_in, True)
return data_weak, file_out
def get_clean(data, exp):
# Remove negative CC values and set to 0
# vrec1 vrec2 vrec1_nmda vrec2_nmda
data['vrec1'] = np.where(data['vrec1'] > 0, 0, data['vrec1'])
data['vrec2'] = np.where(data['vrec2'] > 0, 0, data['vrec2'])
data['vrec1_nmda'] = np.where(data['vrec1_nmda'] > 0, 0,
data['vrec1_nmda'])
data['vrec2_nmda'] = np.where(data['vrec2_nmda'] > 0, 0,
data['vrec2_nmda'])
file_out = exp.name + '_coupling_clean.csv'
pfiles.save_csv(data, file_out, exp.dir_in, True)
return data
def get_measures(list_desc, list_stats, exp, dir_out):
# Get Desc by measure for all genotypes
# Routin for if one of the desc is empty
i = 0
list_measures = list()
for measure in exp.measures:
df_measure = pd.DataFrame()
df_measure[measure] = list_stats # This will be the index column
for desc in list_desc: # one per group from the above desc stats section
df_measure[exp.groups[i]] = desc[measure]
i = i + 1
list_measures.append(df_measure)
file_out = exp.name + '_coupling_' + measure + '_desc.csv'
pfiles.save_csv(df_measure, file_out, dir_out, False)
i = 0
return list_measures
def get_anova(exp, list_groups, measure):
groups_anova = list()
i = 0
while i < len(list_groups):
groups_anova.append(list_groups[i])
i += 1
list_anova = list()
total_n = 0
for df_group in groups_anova:
new_measure = df_group[measure].dropna()
list_anova.append(new_measure)
total_n = total_n + len(new_measure)
# One-way Anova with strain as the independent factor and 3 groups
if len(groups_anova) == 3:
f_value, p_anova = stats.f_oneway(list_anova[0], list_anova[1],
list_anova[2])
elif len(groups_anova) == 4:
f_value, p_anova = stats.f_oneway(list_anova[0], list_anova[1],
list_anova[2], list_anova[3])
freedom1 = len(groups_anova)
freedom2 = total_n - len(groups_anova)
df_anova = pd.DataFrame({
'measure': [measure],
'df_within': [freedom1],
'df_between': [freedom2],
'f_value': [f_value],
'p_value': [p_anova]
})
anova_text = 'One-Way ANOVA: F(' + str(freedom1) + ',' + str(
freedom2) + ') = ' + str(round(f_value, 4)) + ', p = ' + str(
round(p_anova, 4))
df_anova['print'] = anova_text
file_out = exp.name + '_coupling_' + measure + '_' + '_anova_' + '.csv'
pfiles.save_csv(df_anova, file_out, exp.dir_anova, False)
return df_anova
print ('Importing data ........')
# Import each CSV in the directory into dataframes (*****alpha-numeric order!******)
list_dfs = pf.get_csv (dir_in)
# Do basic calculations for cc, gj, rin, ect
print('Calculating Descriptive Stats.......')
list_desc = list()
for df in list_dfs:
list_groupdfs = pf.get_groups (exp, df)
if df.iloc[0]['exp'] == 'mEPSC':
exp.dir_out = exp.dir_in + exp.dir_mEPSCs
# Get t-tests for minis
df_t_mEPSC = ps.ttest_minis (exp, list_groupdfs)
file_out = exp.name + '_mEPSC_t-test.csv'
pf.save_csv (df_t_mEPSC, file_out, exp.dir_out, True)
# Get descriptive stats
for df_group in list_groupdfs:
df_desc = ps.get_desc_mini (exp, df_group)
list_desc.append(df_desc)
print('pass: '******'exp'])
elif df.iloc[0]['exp'] == 'mEPSP':
exp.dir_out = exp.dir_in + exp.dir_mEPSPs
# Get t-tests for minis
df_t_mEPSP = ps.ttest_minis (exp, list_groupdfs)
file_out = exp.name + '_mEPSP_t-test.csv'
pf.save_csv (df_t_mEPSP, file_out, exp.dir_out, True)
def get_desc_cc(exp, df_desc, dir_out):
factor = exp.factor1
confidence = exp.conf
if df_desc.empty:
# Make new dataframe for stats
df_return = pd.DataFrame(columns=[
'stat', 'strain', 'id', 'distance', 'conc_nmda', 'vrec1', 'vstim1',
'istim1', 'vrec1_nmda', 'vstim1_nmda', 'istim1_nmda', 'rin1',
'rin2', 'rin1_nmda', 'rin2_nmda', 'vm1', 'vm2', 'vm1_nmda',
'vm2_nmda', 'bins', 'cc', 'cc_nmda', 'xfer', 'xfer_nmda', 'gj',
'gj_nmda', 'cc_change', 'cc_pchange', 'rin_change', 'rin_pchange',
'gj_change', 'gj_pchange', 'vm_change'
])
df_return['stat'] = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.index = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.reset_index(level=df_return.index.names, inplace=True)
df_return = df_return.rename(columns={'index': 'stat'})
else:
group = df_desc.iloc[0][factor]
# Calculations for all measures at once
n = df_desc.groupby(factor).count()
avg = df_desc.groupby(factor).mean()
sd = df_desc.groupby(factor).std()
se = sd / np.sqrt(n.astype('int'))
added = df_desc.groupby(factor).sum()
minimum = df_desc.groupby(factor).min()
maximum = df_desc.groupby(factor).max()
quartile25 = df_desc.groupby(factor).quantile(q=0.25,
axis=0,
numeric_only=True,
interpolation='linear')
quartile75 = df_desc.groupby(factor).quantile(q=0.75,
axis=0,
numeric_only=True,
interpolation='linear')
median = df_desc.groupby(factor).median()
conf = se * sp.stats.t._ppf((1 + confidence) / 2., n.astype('int') - 1)
conf_5 = avg - conf
conf_95 = avg + conf
n = n.reset_index()
avg = avg.reset_index()
sd = sd.reset_index()
added = added.reset_index()
minimum = minimum.reset_index()
maximum = maximum.reset_index()
quartile25 = quartile25.reset_index()
quartile75 = quartile75.reset_index()
median = median.reset_index()
# Make new dataframe for stats
df_return = pd.DataFrame()
df_return = pd.concat([
n, avg, sd, se, added, minimum, maximum, quartile25, quartile75,
median, conf_5, conf_95
],
sort=False,
ignore_index=False)
df_return.index = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.reset_index(level=df_return.index.names, inplace=True)
df_return = df_return.rename(columns={'index': 'stat'})
file_out = exp.name + '_coupling_' + group + '_desc.csv'
pfiles.save_csv(df_return, file_out, dir_out, False)
return df_return
def get_desc_vsteps(exp, df_desc):
factor = exp.factor1
confidence = exp.conf
exp_type = df_desc.iloc[0]['exp']
df_desc = df_desc[[
'genotype', 'lateral', 'holding', 'vm-50', 'vm-40', 'vm-30', 'vm-20',
'vm-10', 'vm0', 'vm+10', 'vm+20', 'vm+30', 'vm+40', 'vm+50', 'vm+60'
]]
if df_desc.empty:
# Make new dataframe for stats
df_return = pd.DataFrame(columns=[
'stat', 'exp', 'genotype', 'lateral', 'vm', 'holding', 'vm-50',
'vm-40', 'vm-30', 'vm-20', 'vm-10', 'vm0', 'vm+10', 'vm+20',
'vm+30', 'vm+40', 'vm+50', 'vm+60'
])
df_return['stat'] = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.index = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.reset_index(level=df_return.index.names, inplace=True)
df_return = df_return.rename(columns={'index': 'stat'})
else:
group = df_desc.iloc[0][factor]
# Calculations for all measures at once
n = df_desc.groupby(factor).count()
avg = df_desc.groupby(factor).mean()
sd = df_desc.groupby(factor).std()
se = df_desc.groupby(factor).sem()
added = df_desc.groupby(factor).sum()
minimum = df_desc.groupby(factor).min()
maximum = df_desc.groupby(factor).max()
quartile25 = df_desc.groupby(factor).quantile(q=0.25,
axis=0,
numeric_only=True,
interpolation='linear')
quartile75 = df_desc.groupby(factor).quantile(q=0.75,
axis=0,
numeric_only=True,
interpolation='linear')
median = df_desc.groupby(factor).median()
conf = se * sp.stats.t._ppf((1 + confidence) / 2., n.astype('int') - 1)
conf_5 = avg - conf
conf_95 = avg + conf
n = n.reset_index()
avg = avg.reset_index()
sd = sd.reset_index()
added = added.reset_index()
minimum = minimum.reset_index()
maximum = maximum.reset_index()
quartile25 = quartile25.reset_index()
quartile75 = quartile75.reset_index()
median = median.reset_index()
# Make new dataframe for stats
df_return = pd.DataFrame()
df_return = pd.concat([
n, avg, sd, se, added, minimum, maximum, quartile25, quartile75,
median, conf_5, conf_95
],
sort=False,
ignore_index=True)
df_return.index = [
'n', 'mean', 'sd', 'se', 'sum', 'min', 'max', 'quart_25',
'quart_75', 'median', 'ci_5', 'ci_95'
]
df_return.reset_index(level=df_return.index.names, inplace=True)
df_return = df_return.rename(columns={'index': 'stat'})
file_out = exp.name + '_' + exp_type + '_' + group + '_desc.csv'
pfiles.save_csv(df_return, file_out, exp.dir_out, False)
return df_return