rnadtu 0.1.6

Package for running the CIDR algorithm implemented in R, in python

rnadtu

A Python package for running the CIDR algorithm (originally implemented in R) for clustering single-cell RNA-seq data using AnnData objects.

Features

• Run the CIDR algorithm from Python using:

  • cidr() — subprocess-based, using temporary CSV files (recommended for large datasets)
  • cidr_non_csv() — subprocess-based, using in-memory buffers (faster but memory intensive and difficult to debug)
  • cidr_rpy2() — uses the rpy2 bridge (direct R integration, less suitable for large datasets)

• Optional configuration options:

  • data_type: default is "raw", can be "cpm" (counts per million)
  • n_cluster: default is None (CIDR calculates the optimal number), or manually specify an positive integer
  • layer: which layer of the AnnData object to use for input
  • Optional output controls:
    • pc: store principal coordinates
    • dissim: store dissimilarity matrix
    • dropout: store dropout matrix
    • save_clusters: (only in cidr_rpy2) choose store cluster labels (default True)

• Results are saved in the AnnData object:

  • obsm[layer + "_cidr_pc"] — principal components
  • obsm[layer + "_cidr_clusters"] — cluster labels (if save_clusters=True)
  • obsp[layer + "_cidr_dissimilarity_matrix"] — pairwise distances
  • uns[layer + "_cidr_variation"], uns[layer + "_cidr_eigenvalues"] — PCA variation
  • uns[layer + "_cidr_dropout_candidates"] — dropout data

• Generates a clustering plot in cidr_plots.pdf

Installation

pip install rnadtu