Usage

Basic usage

The simplest use of triku is with a common pipeline of scanpy. We show the example with the pbmc3k dataset from scanpy:

import scanpy as sc
import triku as tk

pbmc = sc.datasets.pbmc3k()

sc.pp.filter_cells(pbmc, min_genes=50)
sc.pp.filter_genes(pbmc, min_cells=10)
sc.pp.log1p(pbmc)

sc.pp.pca(pbmc)
sc.pp.neighbors(pbmc)

tk.tl.triku(pbmc)

This is a basic preprocessing of a dataset. You can run triku either after or before sc.pp.log1p. It usually works better after log transformation.

After running triku, results are stored in adata.var (triku_distance, highly_variable), and in adata.uns['triku_params'][None]. Also, adata.obsm['X_triku'] contains the matrix with the selected features. This is relevant to calculate the new round of neighbors or to do PCA again with that data.

Advanced usage

When using triku, there are more some parameters that can be changed. All of them can be found at the API Reference.

  • n_features: The number of features to be selected. For instance, tk.tl.triku(adata, n_features=500) would select the first 500 features.

  • use_raw: Uses counts from adata.raw. This, for instance, can be used to select non log-transformed counts. This can be set as tk.tl.triku(adata, use_raw=True).

  • name: Saves the results with a custom name. For instance, if the name is sample, then the results would be stored in adata.var['triku_distance_sample'], and in adata.uns['triku_params']['sample'].