tomodrgn.analysis.load_dataframe

load_dataframe(*, z: ndarray | None = None, pc: ndarray | None = None, tsne_emb: ndarray | None = None, umap_emb: ndarray | None = None, euler: ndarray | None = None, trans: ndarray | None = None, labels: ndarray | None = None, **kwargs) → DataFrame

Merge known types of numpy arrays into a single pandas dataframe for downstream analysis. Only supplied key word arguments will be added as columns to the dataframe.

Parameters:

• z – array of latent embeddings, shape (nptcls, zdim)

• pc – array of PCA-transformed latent embeddings, shape (nptcls, PC-dim)

• tsne_emb – array of t-SNE-transformed latent embeddings, shape (nptcls, tSNE-dim)

• umap_emb – array of UMAP-transformed latent embeddings, shape (nptcls, UMAP-dim)

• euler – array of Euler angles, shape (nptcls, 3)

• trans – array of translation vectors, shape (nptcls, 2)

• labels – array of labels, shape (nptcls, )

• kwargs – key:value pairs of additional column_name: numpy array of values to add as additional columns to the dataframe

Returns: