def plot_summary_stat_boxplots_by_exp_groups(
arr_df, arr_summary_stats, sample_inds=None, fig_path=None, fig_prefix=None, fig_size=(11, 11)
):
""""
Plot boxplots of summary stats by experimental groups
Parameters:
----------
arr_df: pandas.DataFrame
dataframe of array data
arr_summary_stats: list
list of array summary statistics for which boxplots are to be generated.
sample_inds: [bool array | None (default)]
defines which samples to use in boxplots (allows selecting subset).
If set to None (default) will use all samples.
fig_path: [string | None (default)]
if set, will save figures in fig_path. if None (default) figures are not saved.
fig_prefix: [string | None]
if a subset of the data is plotted, allows specifying a figure name prefix
for saving to files.
fig_size : 2-tuple of integers default set to (11, 11)
(width, height) of figure
"""
if sample_inds is None:
sample_inds = [True] * arr_df.shape[0]
if fig_prefix is None:
fig_prefix = ""
for assay in arr_summary_stats:
f = plt.figure()
f.set_size_inches(fig_size)
mbp.myboxplot_by_labels(arr_df[sample_inds][assay], arr_df[sample_inds]["group"])
plt.title("".join([fig_prefix, " ", assay, "_responses"]))
plt.xlabel("group #")
# save to file only if save_flag is on
if fig_path is not None:
f.set_tight_layout(True)
filename = "".join([fig_path, fig_prefix, assay, "_boxplots_by_groups.png"])
f.savefig(filename, dpi=200)
def plot_summary_stat_boxplots_by_clusters(
arr_df, clusters, prot_names, arr_summary_stats, sample_inds=None, fig_prefix=None, fig_path=None, fig_size=(11, 11)
):
"""
Plot boxplots of summary stats by clusters
Parameters:
----------
arr_df: pandas.DataFrame
dataframe of array data
clusters: dictionary
cluster assignment of each datapoint indexed by ind_dict.keys()
prot_names: list
list of strings of protein antigens from which peptides are on the array_data_filename
arr_summary_stats: list
list of array summary statistics for which boxplots are to be generated.
sample_inds: [bool array | None]
boolean array specifying which samples to use. If None (default) will plot all.
fig_path: [string | None (default)]
if set, will save figures in fig_path. if None (default) figures are not saved.
fig_size : 2-tuple of integers default set to (11, 11)
(width, height) of figure
"""
if sample_inds is None:
sample_inds = [True] * arr_df.shape[0]
for p in prot_names:
for assay in arr_summary_stats:
f = plt.figure()
f.set_size_inches(fig_size)
mbp.myboxplot_by_labels(arr_df[sample_inds][assay], clusters[p])
plt.title("".join([p, " clusters ", assay]))
plt.xlabel("Cluster #")
num_clusters = len(np.unique(clusters[p]))
# save to file only if save_flag is on
if fig_path is not None:
filename = "".join(
[fig_path, fig_prefix, p, "_", assay, "_boxplots_by_clusters_n_", str(num_clusters), ".png"]
)
f.savefig(filename, dpi=200)
f.set_tight_layout(True)
f.set_size_inches(18, 11)
for i, p in enumerate(['SHA_ha', 'SHA_na']):
axarr[i].plot(np.arange(len(ind_dict[p])), bg_df[ind_dict[p]].T)
axarr[i].set_title(p + " BSA responses " + str(i+1) + " (n = " + str(len(bg_df.shape[0])) + ")")
# axarr[i].set_yticks([])
filename = "".join([FIG_PATH, p, "_BSA_responses.png"])
f.savefig(filename, dpi=20)
# plot boxplots of all clusters
for p in ['SHA_ha', 'SHA_na']:
for assay in assays:
f = figure()
f.set_size_inches(18, 11)
mbp.myboxplot_by_labels(arr_df[post_inds][assay], clusters[p])
plt.title("".join([p, " clusters ", assay]))
plt.xlabel('Cluster #')
filename = "".join([FIG_PATH, p, "_", assay, "_boxplots_by_clusters_n_", str(num_clusters), ".png"])
f.savefig(filename, dpi=200)
# plot boxplots of all groups:
for assay in assays + arr_summary_stats:
for t in time_dict.keys():
f = figure()
f.set_size_inches(18, 11)
mbp.myboxplot_by_labels(arr_df[time_dict[t]][assay], arr_df[time_dict[t]]['group'])
plt.title("".join([t, "_", assay, "_responses"]))
plt.xlabel('group #')
filename = "".join([FIG_PATH, t, "_", assay, "_boxplots_by_groups.png"])
f.savefig(filename, dpi=200)
for p in ["SHA_ha", "SHA_na"]:
f, axarr = plt.subplots(1)
f.set_tight_layout(True)
f.set_size_inches(18, 11)
sch.dendrogram(Z_struct[p], color_threshold=np.inf, labels=arr_df.index, orientation="left")
axarr.set_title(p)
filename = "".join([FIG_PATH, p, "_dendrograms.png"])
f.savefig(filename, dpi=200)
# plot boxplots of all clusters
for p in ["SHA_ha", "SHA_na"]:
for sum_stat in arr_summary_stats:
f = plt.figure()
f.set_size_inches(18, 11)
mbp.myboxplot_by_labels(arr_df[sum_stat], clusters[p])
plt.title("".join([p, " clusters ", sum_stat]))
plt.xlabel("Cluster #")
filename = "".join([FIG_PATH, p, "_", sum_stat, "_boxplots_by_clusters_n_", str(num_clusters), ".png"])
f.savefig(filename, dpi=200)
# # plot boxplots of all groups:
# for assay in assays + arr_summary_stats:
# for t in time_dict.keys():
# f = figure()
# f.set_size_inches(18, 11)
# mbp.myboxplot_by_labels(arr_df[time_dict[t]][assay], arr_df[time_dict[t]]['group'])
# plt.title("".join([t, "_", assay, "_responses"]))
# plt.xlabel('group #')
# filename = "".join([FIG_PATH, t, "_", assay, "_boxplots_by_groups.png"])
# f.savefig(filename, dpi=200)
diff_df.drop("modified", axis=1, inplace=True)
print(diff_df)
diff_df.to_csv(path_or_buf=filename, sep="\t")
# now translate indices into cluster inds: - not used since code modified to include tuple of indices into dataframe as index!
# c_inds = np.unravel_index(sig_df_HA.index,(num_clusters, num_clusters-1))
print("Vic HA clusters with significant differences in HAI or NT assays")
print(sig_df_HA)
# plot boxplots of all clusters
for p in ["Vic_HA", "Vic_NA"]:
for assay in Vic_assays:
f = figure()
mbp.myboxplot_by_labels(df[assay], clusters[p])
plt.title("".join([p, " clusters ", assay_strs[assay]]))
plt.xlabel("Cluster #")
filename = "".join([FIG_PATH, p, "_", assay_strs[assay], "_boxplots_by_clusters_n_", str(num_clusters), ".png"])
f.savefig(filename, dpi=200)
# Plot figures for a given clustering solution - currently only performed for the H3N2 victoria strain:
for p in ["Vic_HA", "Vic_NA"]:
f, axarr = plt.subplots(num_clusters, 1)
f.set_tight_layout(True)
f.set_size_inches(18, 11)
# plot clusters
for i in np.arange(num_clusters):
axarr[i].plot(np.arange(len(ind_dict[p])), df[ind_dict[p]].loc[clusters[p] == i + 1].T)
# plot boxplots of all groups:
# only plot for PBS, Vac, and AS03:
curr_df = arr_df[(arr_df.group != 'WT_post_MF59') & (arr_df.group != 'Ob_post_MF59') &
(arr_df.group != 'WT_pre_MF59') & (arr_df.group != 'Ob_pre_MF59')]
curr_time_dict = {}
curr_time_dict['Pre'] = curr_df.group.str.contains('pre')
curr_time_dict['Post'] = curr_df.group.str.contains('post')
for assay in arr_summary_stats:
for t in ['Post']: # time_dict.keys():
f, axarr = plt.subplots(1)
f.set_tight_layout(True)
mbp.myboxplot_by_labels(curr_df[curr_time_dict[t]][assay], curr_df[curr_time_dict[t]]['group'])
axarr.set_title("".join([t, " ", assay.replace('_', ' '), " responses"]), fontsize=16)
axarr.tick_params(axis='both', which='major', labelsize=14)
axarr.set_yscale('log')
filename = "".join([FIG_PATH, t, "_", assay, "_boxplots_by_groups.png"])
f.savefig(filename, dpi=200)
filename = "".join([FIG_PATH, t, "_", assay, "_boxplots_by_groups.eps"])
f.savefig(filename, dpi=1000)
# Figure 4, 5 and 6 - clustering dendrograms, median responses and summary stats of WT vs. Obese for each group:
for a in ['Vac', 'AS03']:
curr_inds = group_inds['Ob_post_' + a].append(group_inds['WT_post_' + a])
amp.plot_clustering_dendrograms(Z_struct=Z_struct[a], prot_names=['SHA_ha'], labels=arr_df.loc[curr_inds].group, fig_prefix=a + '_', fig_path=FIG_PATH)
