aaanalysis.AAclust.comp_correlation

static AAclust.comp_correlation(X, labels=None, X_ref=None, labels_ref=None, names=None, names_ref=None)

Computes the Pearson correlation of given data with reference data.

Parameters:

• X (array-like, shape (n_samples, n_features)) – Feature matrix. Rows typically correspond to scales and columns to amino acids.

• labels (array-like, shape (n_samples,)) – Cluster labels for each sample in X.

• X_ref (array-like, shape (n_samples, n_features)) – Feature matrix of reference data. If given, samples of X are compared with samples of X_ref.

• labels_ref (array-like, shape (n_samples_ref,)) – Cluster labels for each sample in X_ref.

• names (list of str, optional) – List of sample names corresponding to X.

• names_ref (list of str, optional) – List of sample names corresponding to X_ref.

Returns:

• df_corr (pd.DataFrame, shape (n_samples, n_samples or n_samples_ref)) – DataFrame with correlation either for each pair in X of shape (n_samples, n_samples) or for each pair between X and X_ref of shape (n_samples, n_samples_ref).

• labels_sorted (array-like, shape (n_samples_ref,)) – Cluster labels for each sample and sorted as in df_corr.

Notes

• Rows will be sorted in ascending order of labels.

• Columns will be sorted in ascending order of labels or labels_ref if given.

• Labels are replaced by respective names if given.

See also

• pandas.DataFrame.corr() used to compute the correlation.

Examples

To showcase the AAclust().correlation() method, we create an example dataset:

import aaanalysis as aa
import seaborn as sns
import matplotlib.pyplot as plt
aa.options["verbose"] = False
# Obtain example scale dataset
df_scales = aa.load_scales(unclassified_out=True).T.sample(50).T
df_cat = aa.load_scales(name="scales_cat")
dict_scale_name = dict(zip(df_cat["scale_id"], df_cat["subcategory"]))
names = [dict_scale_name[s] for s in list(df_scales)]
X = df_scales.T
# Fit AAclust model and retrieve labels
aac = aa.AAclust()
labels = aac.fit(X, n_clusters=10).labels_

We now use the AAclust().comp_correlation() method to compute the pairwise Pearson correlation across all scales provided as feature matrix X and sorted based on their clustering labels:

df_corr, labels_sorted = aac.comp_correlation(X=X, labels=labels)
# Visualize pairwise correlation by heatmap
sns.heatmap(df_corr, vmax=1, vmin=-1, cmap="viridis", cbar_kws=dict(label="Pearson correlation"))
plt.show()

The correlation between the scales and the obtained cluster centers can be computing by providing additionally their feature matrix X_ref and labels labels_ref:

X_ref, labels_ref = aac.comp_medoids(X, labels=labels)
df_corr, _ = aac.comp_correlation(X=X, labels=labels, X_ref=X_ref, labels_ref=labels_ref)
# Visualize correlation between scales and cluster centers
sns.heatmap(df_corr, vmax=1, vmin=-1, cmap="viridis", cbar_kws=dict(label="Pearson correlation"))
plt.show()

List of names can be provided for all scales (from X) and the reference scales using the names and names_ref parameters:

# Obtain names of clusters
cluster_names = aac.name_clusters(X, labels=labels, names=names, shorten_names=False)
dict_cluster = dict(zip(labels, names))
names_ref = [dict_cluster[i] for i in labels_ref]
# Creat correlation DataFrane including names
df_corr, _ = aac.comp_correlation(X=X, labels=labels, X_ref=X_ref, labels_ref=labels_ref, names=names, names_ref=names_ref)
# Visualize correlation between scales and cluster centers
sns.heatmap(df_corr, vmax=1, vmin=-1, cmap="viridis", cbar_kws=dict(label="Pearson correlation"))
plt.show()