Duplicate code between PCA.py and plot_PCA.py

[AshlinHarris]

Some code for filtering, analysis, and plotting is duplicated between src/PCA.py and src/plot_PCA.py.

[arisp99]

In particular, the duplicate code is in the following lines:

MIPTools/src/PCA.py

Lines 107 to 163 in 6035a73

	pca = decomposition.PCA(n_components=n_components)
	pca.fit(variant_table)
	X = pca.transform(variant_table)
	sample_names = variant_table.index.tolist()
	if (meta_data is not None) and (hue_column is not None):
	sites = meta_data.set_index("Sample ID").loc[sample_names,
	hue_column].values
	unique_sites = set(sites)
	site_color_dict = dict(zip(unique_sites, cycle(all_colors)))
	else:
	sites = np.ones(len(sample_names))
	site_color_dict = {1: "k"}
	fig, axes = plt.subplots(3, 1)
	ax = axes[0]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 0], X[ctry_mask, 1],
	c=color, label=ctry, s=scatter_size)
	if meta_data is not None:
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC1 (%0.1f%%)" % pca.explained_variance_[0],
	fontsize=fontsize)
	ax.set_ylabel("PC2 (%0.1f%%)" % pca.explained_variance_[1],
	fontsize=fontsize)
	ax = axes[1]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 1], X[ctry_mask, 2],
	c=color, label=ctry, s=scatter_size)
	if meta_data is not None:
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC2 (%0.1f%%)" % pca.explained_variance_[1],
	fontsize=fontsize)
	ax.set_ylabel("PC3 (%0.1f%%)" % pca.explained_variance_[2],
	fontsize=fontsize)
	ax = axes[2]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 0], X[ctry_mask, 2],
	c=color, label=ctry, s=scatter_size)
	if meta_data is not None:
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC1 (%0.1f%%)" % pca.explained_variance_[0],
	fontsize=fontsize)
	ax.set_ylabel("PC3 (%0.1f%%)" % pca.explained_variance_[2],
	fontsize=fontsize)
	fig.set_size_inches(*fig_size)
	fig.set_dpi(fig_dpi)

and:

MIPTools/src/plot_PCA.py

Lines 46 to 93 in 6035a73

	pca = decomposition.PCA(n_components=3)
	pca.fit(filled_table.T)
	X = pca.transform(filled_table.T)
	sample_names = filled_table.columns.tolist()
	if (meta_data is not None) and (hue_column is not None):
	sites = meta_data.set_index("Sample ID").loc[
	sample_names, hue_column].values
	unique_sites = set(sites)
	site_color_dict = dict(zip(unique_sites, cycle(all_colors)))
	else:
	sites = np.ones(len(sample_names))
	site_color_dict = {1: "k"}
	fig, axes = plt.subplots(3, 1)
	ax = axes[0]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 0], X[ctry_mask, 1],
	c=color, label=ctry, s=10)
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC1 (%0.1f%%)" % pca.explained_variance_[0], fontsize=16)
	ax.set_ylabel("PC2 (%0.1f%%)" % pca.explained_variance_[1], fontsize=16)
	ax = axes[1]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 1], X[ctry_mask, 2],
	c=color, label=ctry, s=15)
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC2 (%0.1f%%)" % pca.explained_variance_[1], fontsize=16)
	ax.set_ylabel("PC3 (%0.1f%%)" % pca.explained_variance_[2], fontsize=16)
	ax = axes[2]
	for ctry, color in site_color_dict.items():
	ctry_mask = sites == ctry
	if len(X[ctry_mask, 0]) > 0:
	ax.scatter(X[ctry_mask, 0], X[ctry_mask, 2],
	c=color, label=ctry, s=15)
	ax.legend()
	ax.set_xticklabels([])
	ax.set_yticklabels([])
	ax.set_xlabel("PC1 (%0.1f%%)" % pca.explained_variance_[0], fontsize=16)
	ax.set_ylabel("PC3 (%0.1f%%)" % pca.explained_variance_[2], fontsize=16)
	fig.set_size_inches(4, 12)
	fig.set_dpi(150)