Scanpy API

Explored by:

HUANG, HELEN,MATTHEWS, JAMIE EMOTO,PERRIE, JONATHAN VI-(Bioinformatics 201, Winter-2022, UCLA)

img

Description:

Scanpy is a general purpose API for various components of the single-cell analysis pipeline; these include:

  1. Filtering
  2. Low-dimensional Embeddings
  3. Clustering
  4. Marker gene identification
  5. Plotting

It can be accessed via the website documentation, and the original paper can be found on Genome Biology at this link.

Tutorial:

we created a simple tutorial covering some of these functions for a scRNA-seq dataset, which can be accessed at this link . The tutorial has the following structure:

  1. Load data
  2. Filter
  3. t-SNE vs UMAP computation and plotting
  4. Cluster into cell types and plotting
  5. Identify genes important for discriminating into groups

Use cases:

we present two cases centered around analyzing rare cell types.

  1. Task: Given a scRNA-seq data set with rare cell types, identify the marker genes Background: To better understand rare cell types, we can identify genes that are distinct in their gene expression profile Goal: Identify marker genes for rare cell types Method: API pipeline with some downsampling to balance class sizes Outcome: Marker genes for rare cell types

  2. Task: Given a scRNA-seq data set with rare cell types, construct a trajectory mapping the relationships between all cell types Background: To better understand rare cell types, we can look at their ontological frame with respect to other cell types mapped in some embedding space Goal: Append rare cell types to lineage tree of more common cell types Method: API pipeline with a separate projection of rare cell types to embedding after initial trajectory construction Outcome: Cell type trajectory