Beginner tutorial: QC and first UMAP

Goal

Load an AnnData object, run basic QC-oriented helpers, and generate a first UMAP.

Inputs

  • example.h5ad with an expression matrix and common obs QC columns.

Steps

import scanpy as sc
import scanpy_plus as sp

adata = sc.read_h5ad("example.h5ad")
adata.obs["is_outlier"] = sp.is_outlier(adata, metric="total_counts", nmads=5)

adata = sp.pp.assign_sex(adata)
adata = sp.tl.tryumap(adata)

sp.pl.umap_allobs(adata, wanted_obs=["leiden", "sample"], cmap="viridis")

Expected output and interpretation

  • adata.obs['is_outlier'] flags potential QC outliers.

  • adata.obs gets sex-assignment annotations.

  • UMAP visual output helps confirm whether major groups separate as expected.